EXCELLENCE
Specialised Engineering for the Mining Industry
www.PatersonCooke.com
Tailings & Mine Waste
Long-distance Pipelines
Mine Backfill
Mineral Processing
Mine Services
Bulk Water Supply
Dredging & Marine Pipelines
Laboratory & Test Work
Engineering Services
Tailings & Mine Waste
As leading experts, our holistic solutions balance many competing requirements, such as geotechnical stability, geochemical stability, water conservation, social license to operate, people at risk, emissions, operational risks, and cost. Our solutions are site specific and focus on each of the processing, transportation, and deposition challenges, as well as the client’s needs.
Long-distance Pipelines
With over 30 years of experience designing overland pipeline systems, we are global leaders in this field. Our expertise includes conventional tailings pipelines, thickened and paste tailings disposal systems, and long-distance ore and concentrate pipelines.
Mine Backfill
Our engineers have solved some of the most complex backfill challenges while testing and designing hydraulic, paste and cemented rock fill (CRF) projects all over the world. We remain leaders in the design, execution, and ongoing management support of mine backfill systems.
The Southern African Institute of Mining and Metallurgy
OFFICE BEARERS AND COUNCIL FOR THE 2024/2025 SESSION
President E. Matinde
President Elect
G.R. Lane
Senior Vice President
T.M. Mmola
Junior Vice President
M.H. Solomon
Incoming Junior Vice President
S.J. Ntsoelengoe
Immediate Past President
W.C. Joughin
Honorary Treasurer
W.C. Joughin
Ordinary Members on Council
W. Broodryk M.C. Munroe
Z. Fakhraei S.M. Naik
B. Genc G. Njowa
K.M. Letsoalo S.M. Rupprecht
S.B. Madolo A.T. van Zyl
M.A. Mello E.J. Walls
K. Mosebi
Co-opted Council Members
A.D. Coetzee
L.T. Masutha
Past Presidents Serving on Council
N.A. Barcza C. Musingwini
R.D. Beck S. Ndlovu
Z. Botha J.L. Porter
V.G. Duke M.H. Rogers
I.J. Geldenhuys G.L. Smith
R.T. Jones
G.R. Lane – TP Mining Chairperson
Z. Botha – TP Metallurgy Chairperson
K.W. Banda – YPC Chairperson
C.T. Chijara – YPC Vice Chairperson
Branch Chairpersons
Botswana K. Mosebi
DRC K.T. Kekana (Interim Chairperson)
Johannesburg N. Rampersad
Limpopo M.S. Zulu
Namibia T. Aipanda
Northern Cape Vacant
North West Vacant
Pretoria P.G.H. Pistorius
Western Cape Vacant
Zambia N.M. Kazembe
Zimbabwe L. Shamu
Zululand Vacant
PAST PRESIDENTS
*Deceased
* W. Bettel (1894–1895)
* A.F. Crosse (1895–1896)
* W.R. Feldtmann (1896–1897)
* C. Butters (1897–1898)
* J. Loevy (1898–1899)
* J.R. Williams (1899–1903)
* S.H. Pearce (1903–1904)
* W.A. Caldecott (1904–1905)
* W. Cullen (1905–1906)
* E.H. Johnson (1906–1907)
* J. Yates (1907–1908)
* R.G. Bevington (1908–1909)
* A. McA. Johnston (1909–1910)
* J. Moir (1910–1911)
* C.B. Saner (1911–1912)
* W.R. Dowling (1912–1913)
* A. Richardson (1913–1914)
* G.H. Stanley (1914–1915)
* J.E. Thomas (1915–1916)
* J.A. Wilkinson (1916–1917)
* G. Hildick-Smith (1917–1918)
* H.S. Meyer (1918–1919)
* J. Gray (1919–1920)
* J. Chilton (1920–1921)
* F. Wartenweiler (1921–1922)
* G.A. Watermeyer (1922–1923)
* F.W. Watson (1923–1924)
* C.J. Gray (1924–1925)
* H.A. White (1925–1926)
* H.R. Adam (1926–1927)
* Sir Robert Kotze (1927–1928)
* J.A. Woodburn (1928–1929)
* H. Pirow (1929–1930)
* J. Henderson (1930–1931)
* A. King (1931–1932)
* V. Nimmo-Dewar (1932–1933)
* P.N. Lategan (1933–1934)
* E.C. Ranson (1934–1935)
* R.A. Flugge-De-Smidt (1935–1936)
* T.K. Prentice (1936–1937)
* R.S.G. Stokes (1937–1938)
* P.E. Hall (1938–1939)
* E.H.A. Joseph (1939–1940)
* J.H. Dobson (1940–1941)
* Theo Meyer (1941–1942)
* John V. Muller (1942–1943)
* C. Biccard Jeppe (1943–1944)
* P.J. Louis Bok (1944–1945)
* J.T. McIntyre (1945–1946)
* M. Falcon (1946–1947)
* A. Clemens (1947–1948)
* F.G. Hill (1948–1949)
* O.A.E. Jackson (1949–1950)
* W.E. Gooday (1950–1951)
* C.J. Irving (1951–1952)
* D.D. Stitt (1952–1953)
* M.C.G. Meyer (1953–1954)
* L.A. Bushell (1954–1955)
* H. Britten (1955–1956)
* Wm. Bleloch (1956–1957)
* H. Simon (1957–1958)
* M. Barcza (1958–1959)
* R.J. Adamson (1959–1960)
* W.S. Findlay (1960–1961)
* D.G. Maxwell (1961–1962)
* J. de V. Lambrechts (1962–1963)
* J.F. Reid (1963–1964)
* D.M. Jamieson (1964–1965)
* H.E. Cross (1965–1966)
* D. Gordon Jones (1966–1967)
* P. Lambooy (1967–1968)
* R.C.J. Goode (1968–1969)
* J.K.E. Douglas (1969–1970)
* V.C. Robinson (1970–1971)
* D.D. Howat (1971–1972)
* J.P. Hugo (1972–1973)
* P.W.J. van Rensburg (1973–1974)
* R.P. Plewman (1974–1975)
* R.E. Robinson (1975–1976)
* M.D.G. Salamon (1976–1977)
* P.A. Von Wielligh (1977–1978)
* M.G. Atmore (1978–1979)
* D.A. Viljoen (1979–1980)
* P.R. Jochens (1980–1981)
* G.Y. Nisbet (1981–1982)
A.N. Brown (1982–1983)
* R.P. King (1983–1984)
J.D. Austin (1984–1985)
* H.E. James (1985–1986)
H. Wagner (1986–1987)
* B.C. Alberts (1987–1988)
* C.E. Fivaz (1988–1989)
* O.K.H. Steffen (1989–1990)
* H.G. Mosenthal (1990–1991)
R.D. Beck (1991–1992)
* J.P. Hoffman (1992–1993)
* H. Scott-Russell (1993–1994)
J.A. Cruise (1994–1995)
D.A.J. Ross-Watt (1995–1996)
N.A. Barcza (1996–1997)
* R.P. Mohring (1997–1998)
J.R. Dixon (1998–1999)
M.H. Rogers (1999–2000)
L.A. Cramer (2000–2001)
* A.A.B. Douglas (2001–2002)
* S.J. Ramokgopa (2002-2003)
T.R. Stacey (2003–2004)
F.M.G. Egerton (2004–2005)
W.H. van Niekerk (2005–2006)
R.P.H. Willis (2006–2007)
R.G.B. Pickering (2007–2008)
A.M. Garbers-Craig (2008–2009)
J.C. Ngoma (2009–2010)
G.V.R. Landman (2010–2011)
J.N. van der Merwe (2011–2012)
G.L. Smith (2012–2013)
M. Dworzanowski (2013–2014)
J.L. Porter (2014–2015)
R.T. Jones (2015–2016)
C. Musingwini (2016–2017)
S. Ndlovu (2017–2018)
A.S. Macfarlane (2018–2019)
M.I. Mthenjane (2019–2020)
V.G. Duke (2020–2021)
I.J. Geldenhuys (2021–2022)
Z. Botha (2022-2023)
W.C. Joughin (2023-2024)
Editorial Board
S.O. Bada
R.D. Beck
P. den Hoed
I.M. Dikgwatlhe
M. Erwee
B. Genc
R.T. Jones
W.C. Joughin
A.J. Kinghorn
D.E.P. Klenam
D.F. Malan
D. Morris
C. Musingwini
P.N. Neingo
S.S. Nyoni
M. Onifade
M. Phasha
P. Pistorius
P. Radcliffe
N. Rampersad
Q.G. Reynolds
I. Robinson
S.M. Rupprecht
T.R. Stacey
International Advisory Board members
R. Dimitrakopolous
R. Mitra
S. Ndlovu
A.J.S. Spearing
E. Topal
D. Tudor
F. Uahengo
D. Vogt
Editor /Chairperson of the Editorial Board
R.M.S. Falcon
Typeset and Published by
The Southern African Institute of Mining and Metallurgy
PostNet Suite #212 Private Bag X31 Saxonwold, 2132
E-mail: journal@saimm.co.za
Printed by Camera Press, Johannesburg
Advertising Representative
Barbara Spence
Avenue Advertising
VOLUME 125 NO. 5 MAY 2025
Contents
Journal Comment: Think small by D. Vogt
President’s Corner: Engineers dividend and the African mine of the future by E. Matinde
THE INSTITUTE, AS A BODY, IS NOT RESPONSIBLE FOR THE STATEMENTS AND OPINIONS ADVANCED IN ANY OF ITS PUBLICATIONS.
Copyright© 2025 by The Southern African Institute of Mining and Metallurgy. All rights reserved. Multiple copying of the contents of this publication or parts thereof without permission is in breach of copyright, but permission is hereby given for the copying of titles and abstracts of papers and names of authors. Permission to copy illustrations and short extracts from the text of individual contributions is usually given upon written application to the Institute, provided that the source (and where appropriate, the copyright) is acknowledged. Apart from any fair dealing for the purposes of review or criticism under The Copyright Act no. 98, 1978, Section 12, of the Republic of South Africa, a single copy of an article may be supplied by a library for the purposes of research or private study. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means without the prior permission of the publishers. Multiple copying of the contents of the publication without permission is always illegal. U.S. Copyright Law applicable to users In the U.S.A. The appearance of the statement of copyright at the bottom of the first page of an article appearing in this journal indicates that the copyright holder consents to the making of copies of the article for personal or internal use. This consent is given on condition that the copier pays the stated fee for each copy of a paper beyond that permitted by Section 107 or 108 of the U.S. Copyright Law. The fee is to be paid through the Copyright Clearance Center, Inc., Operations Center, P.O. Box 765, Schenectady, New York 12301, U.S.A. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale.
Honorary Legal Advisers
M H Attorneys Auditors
Genesis Chartered Accountants Secretaries
The Southern African Institute of Mining and Metallurgy 7th Floor, Rosebank Towers, 19 Biermann Avenue, Rosebank, 2196
PostNet Suite #212, Private Bag X31, Saxonwold, 2132 E-mail: journal@saimm.co.za
Telephone (011) 463-7940 . E-mail: barbara@avenue.co.za
ISSN 2225-6253 (print) . ISSN 2411-9717 (online)
PROFESSIONAL TECHNICAL AND SCIENTIFIC PAPERS
Comparison of Denver and Leeds laboratory flotation cells: Effect of particle size and hydrodynamics by
M.P. Tshazi, L.S. Leal Filho, N. Naude
.............................................................................
In this paper, the authors performed experiments in the Denver and Leeds laboratory flotation devices1 at the University of Pretoria using various particle sizes to evaluate the performance of the devices. These findings highlighted the influence of particle size on flotation recovery. For all experiments, the Denver laboratory flotation cell outperformed (defined by higher recovery) the Leeds device, while still operating at a lower impeller speed.
The impact of junior miners on the global supply of high-purity manganese sulfate monohydrate for the electric vehicle battery market by
K. Fichani, L.S. Teseletso, J. Kaavera, T.K. Dintwe, E. Shemang, B.I. Matshediso
In this paper, the high-purity manganese sulphate monohydrate project pipeline and projected production volumes with C1 cash costs were researched for six projects in the southern part of Botswana. The results showed that, while the estimated C1 cash costs would place the K.Hill project as the second highest cost producer, it would nonetheless have healthy profit margins due to the high projected price of high-purity manganese sulphate monohydrate. The study recommended policy options, which if implemented, could further encourage the exploitation and value addition of battery metals in Botswana.
The potential of 4IR technologies to mitigate risk in mine residue management by G.M.
Rohde, C. Bester
The fourth industrial revolution (4IR) is impacting society due to increased interconnectivity, processing speed, and automated technologies. This paper explores whether 4IR technologies have the potential to mitigate risks associated with tailings dams. Experts in the field of tailings dam operational management were surveyed to determine their viewon the risks associated with tailings dams and the potential of 4IR technologies to mitigate these risks. The survey found that a majority of experts believe that certain 4IR technologies have the potential to reduce the risks associated with tailings dam failures.
Improving spatial mine-to-plan compliance at an open pit mine through enhanced short-term mine planning by T.J. Otto, T. Mkhatshwa, T.J. van Heerden, C.H. Cloete
The Sishen open pit iron ore mine (Sishen) strives to continuously improve spatial compliance to the business plan. The enhanced short-term mine planning process focuses on detailed tactical sequence designs per mining pushback, the health of value chain buffers, spatial plan-to-plan reconciliation, and the associated management routines. This led to improved spatial mine-to-plan compliance. These results indicate that the application of short-term mine planning can contribute positively to improving the level of spatial execution against the business plan.
and recovery of
and aluminium
272 by M. Kruger, H. Krieg, D. van der Westhuizen
This study aimed to optimally separate cobalt and aluminium from the spent Fischer-Tropsch synthesis catalyst for potential use in the agricultural sector. It leverages the OLIÒ database to predict metal speciation during solvent extraction to aid with experimental planning. The investigation revealed the pivotal role of the aqueous pH yielding effective separation. Equipment design analysis for a targeted separation efficiency of 87% dictated the necessity of two mixer stages and a settler.
by C. Felix, O. Barron
In 2023, polymer electrolyte membrane fuel cell membrane electrode assemblies accounted for about 62% of the global fuel cell market share and this demand is expected to grow. The membrane electrode assembly is crucial and represents a significant cost to polymer electrolyte membrane fuel cell manufacturing. Roll-to-roll is a continuous coating process with speeds of 10 m/min being achieved. In this work, MicrogravureTM and slot-die coating methods have been adopted and have shown promising outcomes. An overview of the MicrogravureTM and slot-die coating methodologies is provided, and a discussion on preliminary results and challenges are presented.
Separation
cobalt
from spent gas-to-liquid catalysts using CYANEX®
Roll-to-roll coating methods for the manufacture of polymer electrolyte membrane fuel cell membrane electrode assemblies
HYDROGEN
Journal Comment
MThink small
ining is big business. For many commodities, the orebodies are big and the best way to exploit them for maximum profit is on a large scale.
But do large mines mean large equipment? At the moment, the answer is yes. Ramps can only handle a limited number of trucks per day and removing that obstacle is very expensive; more or wider ramps, a larger pit to hold them, and a lower extraction ratio. We deal with the problem of optimising ramps by making our trucks as large as possible. This also means we can manage with fewer drivers and, as is common knowledge, drivers are an ongoing expense.
If we look at air travel, we have seen the same move to larger equipment over time. A route like Johannesburg to London is like a ramp. It cannot take many more aircraft than it already does, because airports are constrained to accept a limited number of aircraft per day. The limit is safety; aircraft cannot land more frequently because of fear of collision. The result is a longterm trend to larger and more fuel-efficient aircraft.
But aviation is changing. We are seeing developments in local flight like the Lilium air-taxi: electric power and vertical take-off enables quiet ‘air taxis’ that can fly directly from your house to the nearby airport, from where you can catch a plane to anywhere.
For both mine haulage and aircraft, a constraint is a driver. While aviation is becoming steadily more automated, it is unlikely that we will see pilotless planes for a while yet, more because we cannot stomach the idea of a computer flying the plane in which we sit than because it is going to be less safe. For mine haulage, we already have automated trucks, and their safety record is better than that of human-driven trucks.
In mining, it is unlikely that anyone is going to switch to many small trucks in a large operation, although there is evidence that their flexibility might make them a more costeffective option. But then again, automating small trucks in a small operation makes sense. For example, a planned mine nearby can only work during the day, due to concerns about noise from its neighbours. In their application, a small, autonomous electric haul truck would be able to operate at night because it is silent, and in this case, would travel downhill loaded and uphill empty, and subsequently may be able to achieve its task without consuming any diesel.
With time, perhaps we will see many smaller trucks in large pits, rather than a few larger trucks, running on electricity rather than diesel, respectively. If we are serious about geometallurgy, we need to handle ore in smaller packages, and just the improvement in grade control could make the switch possible. With the widespread introduction of renewable energy, it also allows mines to lower their diesel bills and be seen to be greener.
At a time of tariffs and uncertainty, anything that can reduce risks and lower costs appears good. General automation of the mining fleet is at the point where it can solve problems and make mines at all scales more efficient, and therefore, more viable when prices collapse.
D. Vogt
President’s Corner
TEngineers dividend and the African mine of the future
he mining industry has experienced massive metamorphic and irreversible structural changes in the recent past. In addition to the recent unpredictable geopolitical conditions, the major challenges affecting the actors in the mining industry are intricately shaped by structural constraints such as geological, technological, and market conditions. Complexity in the geometallurgical properties of the individual ore bodies, for example, is irreversibly shaped by the geological conditions that existed billions of years ago. The grade and mineralisation properties of the ore bodies also have a significant impact on the choice of mining and processing technologies adopted, the economics of production, and the location of the mining operations. In addition, the mining industry also continues to face operational pressures to cut costs and increase productivity, while simultaneously navigating other challenges such as competition for high end skills, and meeting increasingly strict governance requirements and stakeholder expectations.
Due to the nature of mining as a business, it is clear that these inherent challenges are here to stay, and thus, the future of the industry depends on the ability to learn and adapt. The first, and perhaps most important priority for any mining operation, would be to strengthen its operational and cashflow resilience to enable it to weather the obvious challenges, such as geopolitical disruptions and cyclical downturns. Second to strengthening the economic position, the focus of a future looking mining operation would be to elevate its social licence to operate, achieved mostly through long term investments in human capital and environmental, social, governance, and sustainability KPIs. In fact, investing in talent and leadership has been considered a key variable to building a sustainable mining operation, as the mine of the future will be highly automated and will require highly skilled personnel capable of operating sophisticated systems and technologies. Establishing and strengthening economic linkages with other sectors of the economy is also critical to building value chain resilience and mitigating against the cyclical impact of the commodity markets.
In my October 2024 article, I introduced a controversial and yet highly ambitious proposition that critical minerals can result in sustained technological and economic catch-up. My hypothesis still remains unchanged, and I am convinced that it is possible to utilise the vast experience in complex mining systems and technologies on the African continent to build a vibrant manufacturing economy capable of providing value-added products and services to the rest of the world. I also highlighted that the ability to catch up is driven by deliberate efforts to build value-add competencies. Traditionally, the number of science, technology, engineering, and mathematics (STEM) skills active in the economy was used as a proxy measure of technological capabilities, however more recently, interesting terms such as ‘engineer dividend’ are being introduced to broadly describe the nature and quantum of STEM skills that are required to drive and sustain technological innovation.
President’s Corner (continued)
Borrowing from investment economics, the term ‘engineer dividend’ was introduced in a recent Bloomberg article to refer to a phenomenon whereby countries with a large and diverse skilled engineering workforce naturally develop competitive advantage in areas leading to technological advancement and industrial development (Bloomberg, available at https://www.bloomberg.com/opinion/articles/2025-03-24/china-s-engineer-dividendis-paying-off-big-time). Case in mind is China, which, according to Bloomberg, invested intensely in STEM education and managed to increase its number of engineers by close to 12 million in the period between 2000 and 2020. In this context, engineer dividend was described as an internal rate of return on investment in engineering competencies to create a concentrated community of experts and strong network of engineering skills capable of driving innovation in multiple fields. Although dependent on other factors to succeed, the high concentration of multidisciplinary and specialised engineering expertise fosters both competition and collaboration, leading to improved sector productivity and industrial competitiveness.
In conclusion, there is no doubt that resource-rich countries on the continent can leverage on the extensive experience in designing and operating complex mining systems and operations to build a robust manufacturing and value-added services economy. With all conditions remaining the same, would it be plausible to consider the ‘engineer dividend’ concept as a viable skills investment concept to derisk the future of African mining industry and, if so, which areas should we focus on?
E. Matinde President, SAIMM
Affiliation:
1University of Pretoria, South Africa
2Universidade de São Paulo, Brazil
Correspondence to:
M.P. Tshazi
Email: mfesane.tshazi@up.ac.za
Dates:
Received: 25 Aug. 2022
Revised: 29 Nov. 2024
Accepted: 27 Mar. 2025
Published: May 2025
How to cite:
Tshazi, M.P., Leal Filho, L.S., Naude, N. 2025. Comparison of Denver and Leeds laboratory flotation cells: Effect of particle size and hydrodynamics. Journal of the Southern African Institute of Mining and Metallurgy, vol. 125, no. 5, pp. 225–232
DOI ID:
https://doi.org/10.17159/2411-9717/2289/2025
ORCiD:
M.P. Tshazi
http://orcid.org/0009-0001-6389-1547
L.S. Leal Filho
http://orcid.org/0000-0001-8501-1857
N. Naude
http://orcid.org/0000-0002-9615-0243
Comparison of Denver and Leeds laboratory flotation cells: Effect of particle size and hydrodynamics
by M.P. Tshazi1, L.S. Leal Filho², N. Naude1
Abstract
In this paper, the authors performed experiments in the Denver and Leeds laboratory flotation devices1 at the University of Pretoria at various particle sizes to evaluate the performance of the devices. Quartz was used in a single mineral system at discrete sizes fractions, -25 µm, +2545 µm, +45-75 µm, and +75-106 µm. Hydrodynamic analysis, based on dimensionless power and Reynolds numbers, indicated that the Leeds cell required higher power input to achieve comparable flow conditions. Specifically, the Leeds cell exhibited an average power number of 1.03, whereas the Denver cell averaged 0.77 within the same impeller speed range of 1000 rpm–1500 rpm. For comparative flotation performance, impeller speeds were calibrated, resulting in operating speeds of 1200 rpm for the Denver cell and 1400 rpm for the Leeds cell. The cells performed similarly at two coarser-sized fractions. However, some deviations were observed in the finer particle size range. The -25 µm (fine) fraction initially achieved a recovery of only 15%. An additional reagent dosage was required to enhance the recovery of this fraction significantly. These findings highlighted the influence of particle size on flotation recovery. For all experiments, the Denver laboratory flotation cell outperformed (defined by higher recovery) the Leeds device, while still operating at a lower impeller speed. This advantage can be attributed to its impeller-stator design and air dispersion features, effectively overcoming slurry resistance and resulting in superior flotation performance compared to the Leeds cell.
Keywords quartz, particle size, hydrodynamics, impeller speed, recovery
Introduction
Froth flotation
Froth flotation is commonly used to concentrate valuable minerals. This process depends on adequate surface exposure, i.e., the extent of liberation of the mineral, and selective hydrophobicity of the desired mineral(s). Froth flotation is regarded as a complex system, with many influencing factors and interdependent interactions. These factors and interactions are summarised as a flotation system, which is comprised of three categories of components namely: chemistry, equipment, and operation components. Each component is subdivided into many variables, such as particle size, collectors, and airflow, as shown in Figure 1. These flotation variables significantly influence flotation performance. Laboratory flotation tests have made a considerable contribution to industrial process development and troubleshooting (Newcombe et al., 2012; Ross, 2019). The validity of correlation of laboratory batch froth flotation data to an industrial application is historically somewhat poor (Newcombe et al., 2012). This poor correlation has been linked to differences in equipment and operational variables between the laboratory and plant configurations.
A comparison of the Denver and Leeds laboratory cells was conducted over 40 years ago (Liddell, Dunne, 1984). However, the unique characteristics of these cells, like their structure, impeller, and air intake mechanism, were not explored. The study aimed to compare laboratory flotation cells, Denver and Leeds, across varying particle size ranges and hydrodynamic conditions to provide insight into how particle size and cell characteristics influence flotation performance under identical experimental conditions.
1All references to devices in this document refer to those used at the University of Pretoria.
Comparison of Denver and Leeds laboratory flotation cells: Effect of particle size
Figure 2 (left) illustrates a Denver D-12 laboratory flotation cell, which is widely used for conducting laboratory test work. Figure 2 (right) illustrates the Leeds laboratory flotation cell. As introduced by Professor C. Dell of the University of Leeds, this cell aimed to increase reproducibility of results and reduce operatordependent factors. (Liddel, Dunne, 1983). Flotation tests performed in laboratories are usually dependent on the operator because froth scraping is generally done manually. The scraping areas of the cells differ, with the Denver cell having the impeller in the centre, which obstructs scraping, while the Leeds cell has no obstruction. The Leeds cell features a froth crowder-like shape at the rear. The Leeds flotation cell is designed as a complete unit and parts are not easily interchangeable, whereas the Denver allows for much more flexibility as the cell size and/or impeller design can be changed easily without a significant redesign.
One of the biggest differences between the two cells is their impeller design (Figure 3): The Denver design has four equally sized holes from which the air is sheared as it is introduced into the cell; in the Leeds cell, air is introduced just above the impeller to be sheared. Some air is forced into the impeller and sheared to smaller equally sized bubbles, and some air bypasses the impeller to enter the cell. These differences cause different bubble flow patterns.
Mechanisms
In froth flotation, minerals are concentrated based on three main principles: true flotation, entrainment, entrapment, or a combination of these processes (Wills, Finch, 2015). True flotation is governed by three mechanisms: collision, attachment, and detachment. A brief description of true flotation is that a particle needs to collide with an air bubble, the two entities become attached, and then this unit must remain stable as it moves through the cell and into the froth zone until it is discharged via the cell lip. This is a chemically activated process (Wills, Finch, 2015). Entrainment is the result of water-suspended minerals deporting to the concentrate launder (Wang, 2016). The entrainment mechanism is non-selective as particles are not attached to air bubbles but are instead carried upward by the flow of water or by the wake created by bubble-particle aggregates (Smith, Warren, 1989). This mechanism can affect both valuable and gangue mineral particles. Wang et al. (2015) and Zheng et al. (2006) further observed that entrainment typically favours particle sizes below 50 μm, while entrapment occurs when minerals are trapped between air bubbles because of a non-draining froth phase. This mechanism primarily affects coarse particles (> 106 μm) more than fine particles (Zheng
et al., 2006). Like entrainment, entrapment can concentrate both valuable and gangue minerals, leading to reduced selectivity in the flotation process.
The floatability of a mineral is determined by its particle hydrophobicity (water repellence) and the hydrodynamic conditions that promote particle–bubble collisions, adhesion, and transport from the pulp to the froth. High floatability is typically observed in particles with large contact angles and high collection efficiency (Ek), which is expressed by Equation 1: [1]
where Ec is efficiency of particle–bubble collision; Ea is efficiency of adhesion; Ep is efficiency of preservation of the particle–bubble aggregate.
Particle size
In Trahar’s (1981) study, it was highlighted that although particle size was recognised as a crucial factor in froth flotation, its practical advantages in the design and plant operations have been limited. This gap is expected to close to meet the challenges of processing increasingly complex ores. Particle size plays a crucial role in the efficiency of bubble-particle attachment (Rao, 2004), and flotation recovery is significantly influenced by the size of mineral particles (Jameson, 2012). Norori-McCormac et al. (2017) further demonstrated that even small changes in particle size distribution can significantly impact froths stability.
In addition to the influencing factors recorded in Figure 1, the effect of particle size on flotation recovery is highlighted in Figure 4, categorising particle sizes into three regions: fine, intermediate, and coarse. The specific boundaries of these regions depend on the mineral type and its properties. It is generally accepted that there is an ideal size range (intermediate) where recovery is maximised. Consequently, particles outside of a particular range will be compromised when floated and it is difficult to achieve optimum recoveries (Trahar, 1981).
Figure 1—Various factors influencing a flotation system (Klimpel, 1984 )
Figure 2—Photographs of Denver (left) and Leeds (right) laboratory cells at the University of Pretoria
Figure 3—Impeller and stator configurations of the Denver (left and centre) and Leeds (right) laboratory cells
Comparison of Denver and Leeds laboratory flotation cells: Effect of particle size
Pease et al. (2006), expounded on Trahar’s findings, demonstrating the benefits of floating fine particles separately to maximize recovery (Figure 5). The study emphasised the importance of narrow size distribution in flotation operations, suggesting that both fine and coarse materials can be floated optimally if the kinetics associated with each size fraction are fully understood and controlled.
Fine particles suffer from poor collision efficiency, thus resulting in poor overall recovery (Nguyen, 2007; El-Rahiem, 2014). The main factors affecting this are bubble size, energy input, and pulp rheology, and therefore the recovery performance can be tuned by adjusting either one of those inputs. Moreover, due to the larger surface area of fine particles, higher reagent dosages are often required to render them hydrophobic (Pease et al., 2006). On the other hand, coarse particle flotation is decreased by the detachment of particles from bubbles (Wills, Finch, 2015). The main factors affecting this are hydrophobicity and energy input. Coarse particle recovery can be optimised by focusing on these inputs; however, it should be noted that fine and coarse particle recovery often react in an opposing manner (Safari et al., 2016). For instance, increasing energy input will benefit fine particle recovery while decreasing coarse particle recovery, and careful consideration should be used when optimising recovery of a certain size fraction.
Finer particles generally exhibit slower flotation rates compared to larger particles under conditions of low turbulent energy dissipation, as observed by Pyke et al. (2003). As a result Pyke et al. (2003) and Changunda et al. (2012), observed that kinetic rate (k) increases almost linearly with particle size. This phenomenon is related to the efficiency of particle-bubble collision.
Further research by Murhula, Hashan, and Otsuki (2022) demonstrated that flotation recovery is influenced by the interaction between solid concentration, particle size, and mineral type. In quartz flotation, for instance, higher recovery rates are typically achieved with higher solid concentrations and smaller particle size fractions, highlighting the role of entrainment, whereas more dilute pulps tend to enhance bubble loading efficiency while minimising entrainment (Ramlall, 2008).
Hydrodynamics
Hydrodynamics (fluid flow) within the flotation system are primarily influenced by the impeller (Shabalala et al., 2011; Souza Pinto et al., 2018). Hydrodynamic behaviour is typically characterised by hydrodynamic parameters and dimensionless numbers (Souza Pinto et al., 2018). These parameters and numbers account for the complex interactions and behaviours observed in the flotation system, as illustrated in Figure 1. Key dimensionless numbers include the power number (NP), Reynolds number (NRe), and Froude number (NFr), amongst others.
Dimensionless hydrodynamic numbers are valuable for benchmarking the performance of impellers and flotation cells (Rodrigues et al., 2001). Key dimensionless numbers, such as the NRe, NFr, and NP, are derived from operational variables (e.g., impeller rotational speed, volumetric airflow, pulp specific gravity, and dynamic viscosity) or geometric parameters (e.g., impeller diameter). These dimensionless numbers facilitate the analysis of fluid flow characteristics within a flotation cell. Table 1 provides a comprehensive summary of these parameters, including their formulas and typical ranges, enabling researchers to quickly assess fluid dynamics.
Characterisation of flotation hydrodynamics with dimensionless numbers related to impeller characteristics (from Fuerstenau et al.,
(Wang, Liu, 2021)
Describes solids suspension and characterisation of mixing intensity (Nelson, Lelinski, 2000)
Describes
and how the impeller draws power (Tabosa et al., 2016b)
Table 1
2007)
Figure 4—The effect of particle size on flotation recovery (Trahar, 1981)
Figure 5—An illustration of recovery trend as a function of particle size in three regions, modified from Pease et al. (2006)
Comparison of Denver and Leeds laboratory flotation cells: Effect of particle size
where P is power drawn by impeller (W); N is impeller rotational speed (s-1); D is impeller diameter (m); ρ is specific gravity (kg/m3); μ is dynamic viscosity at a given temperature (kg/m.s); g is gravity acceleration (m/s2).
NP represents the net power drawn for both pumping and shearing, quantifying the resistance imposed by the slurry on the impeller blades, also known as the drag coefficient (CD). A higher NP signifies greater fluid displacement (Arbiter, Harris, 1962), which is advantageous for fine particles due to their low collision efficiency. NP varies with the slurry flow regime and is typically plotted against the Reynolds number (NRe), as shown in Figure 6. In laminar flow, NP decreases steadily, whereas in highly turbulent conditions (NRe > 10⁴), commonly observed in mechanical flotation cells, it remains constant (Harris, 1976; Harris, 1986; Leal Filho et al., 2002).
The dynamic viscosity of water at different temperatures is well-documented in literature. At 27°C, it is reported to be 0.8509 × 10-³ Pa·s (Ma et al., 2020). To estimate the dynamic viscosity of a slurry system, the Krieger–Dougherty Equation (5) is commonly used, particularly for monodisperse systems. The maximum packing fraction (φm) generally falls between 0.6 and 0.7, with a value of 0.63 often used for randomly packed spheres. Furthermore, for spherical particles, the intrinsic viscosity ([η]) is typically 2.5, as noted by Abo Dhaheer et al. (2015).
where μ and μ0 are the dynamic viscosities of the slurry and water (Pa s), respectively; φ is volume fraction of the dispersed solid.
Experimental Laboratory flotation cells
[5]
(medium), and +75–106 µm (coarse). Each fraction was separately blended and split with a rotary splitter to ensure that representative samples were used for the flotation tests. The cell operations were characterised in such a way that comparable kinetic results could be obtained. This was achieved by manually calibrating the impeller speeds of the cells, resulting in optimised speeds of 1200 rpm for the Denver cell and 1400 rpm for the Leeds cell. All tests were performed at 17 mass% solids in a 25 g/t Flotigam EDA ether amine collector, with a 2-minute conditioning time. This collector exhibits frothing properties and therefore no frother was added during the experiments. The pulp was adjusted to a pH of 9.5 with the aid of NaOH. An air flow of 2 L/min was employed. Three to four concentrates were collected every minute after manual scraping at a 10-second interval. It was necessary to conduct further tests on the –25 µm fraction at 50 and 175 g/t collector dosages of Flotigam EDA to improve recovery. Consequently, seven concentrates were collected from the higher collector dosed -25 µm fraction. In general, flotation continued until no froth was recoverable. The concentrates were filtered, oven dried, and weighed. All experiments were carried out repeatedly, with a variability limit of up to a maximum of 5% standard deviation.
The power consumption of Denver and Leeds laboratory cells was measured using an Efergy Classic Wireless Energy Monitor. Tests were conducted at impeller speeds between 1000 rpm and 1500 rpm with increments of 1000 rpm. Testing was conducted under three conditions: an empty cell, a cell filled with 3 L of water with an air flow rate of 2 L/min, and a slurry mixture containing approximately 17 mass% quartz solids with an air flow rate of 2 L/min. The power draw was calculated by subtracting the power consumed under loaded conditions (water or slurry) from that recorded in the no-load (empty) condition.
Results and discussion
All tests were conducted using the Denver and Leeds laboratory flotation cells with nominal volumes of 3.5 L, both made of Perspex. Each cell was equipped with a distinct tachometer for adjusting the impeller speed. A centralised flow meter was employed to regulate the flow of air supplied from a compressed dry air cylinder, ensuring consistent and controlled aeration across the experiments. The Denver and Leeds cells had impeller diameters of 0.07 m and 0.074 m, respectively. The effect of particle size distribution was evaluated in the Denver and Leeds laboratory flotation cells, while maintaining all other variables (given in Figure 1) constant.
Experimental conditions and procedures
A 40 kg sample of 99.9% quartz was used. Quartz is classified as a strongly hydrophilic mineral (Wills, Finch, 2015) and therefore requires chemical activation. The material was wet-milled in a rod mill and thereafter wet-sieved to produce four narrow size fractions: −25 µm (fine), +25–45 µm (intermediate), +45–75 µm
This section presents the results and discussions from a comparative study of Denver and Leeds cells at different particle size fractions. The study was conducted using a single mineral system, namely quartz, ensuring consistent mineralogy and resulting in uniform hydrophobicity.
Effect of particle size
The flotation of four particle size fractions was evaluated under optimised conditions. Figure 7 illustrates the average cumulative recoveries over time for the two larger size fractions: (medium) +45–75 µm, and (coarse) +75–106 µm. Both flotation cells performed similarly for the +45–75 µm fraction, achieving recoveries of approximately 80%. Whereas, for the +75–106 µm fraction, the Denver cell recorded a recovery of 74%, outperforming the Leeds cell by approximately 3%. The difference in performance between the two cells for the coarsest fraction could be attributed to the effect of turbulence, which is better dampened in the Denver cell compared to the Leeds cell. As a consequence of turbulence, which can disrupt particle–bubble aggregates, particularly when coarse particles are present, lower flotation efficiency is the result. (Yao et al., 2021). The recovery of both size fractions was completed within 2.5 minutes, which aligns with expectations, since these fractions fall mostly within the ideal flotation range of +25–106 µm (as shown in Figure 5). These recoveries were aided by the formation of stable and voluminous froths during flotation.
The kinetic constants were approximately the same for both cells for flotation of the intermediate (+45−75 µm) and coarse (+75−106 µm) size fractions, although the values for the Denver cell were slightly higher, aligning with the same trend as the recovery rates.
Figure 6—Drag coefficient (CD) as a function of flow regime (NRe) (Libii, 2010)
Comparison of Denver and Leeds laboratory flotation cells: Effect of particle size
Figure 7—Comparison of Denver (dotted lines) and Leeds (solid lines) cells for kinetics curves for flotation of the medium (+45−75 µm and coarse (+75−106 µm) size fractions of silica, with corresponding rate constant in the legend label
Figure 8 presents the results for the smaller size fractions: fines (−25 µm) and intermediate (+25–45 µm). Recovery of the fine fraction (−25 µm) was very low for both cells, though the Denver cell achieved a slightly higher recovery of 15%. To support this, the Denver cell demonstrated nearly double the kinetic rate (0.32 min-¹) compared to the Leeds cell (0.18 min-¹), suggesting that the Denver impeller enhanced collision efficiency (Ec). The low recovery of the finest particles (−25 µm) can be largely attributed to the low collector dosage (25 g/t), which significantly hindered their flotation performance. Additionally, fines typically have a higher specific surface area (cm²/g), leading to lower collision efficiencies between particles and bubbles. As Pease et al. (2006) showed, fines can achieve good flotation performance, but only when flotation conditions are specifically tailored to treat this fraction in a narrow size distribution range.
For the intermediate size fraction (+25–45 µm), the Denver cell outperformed the Leeds cell, with recoveries of 74% and 63%, respectively. This suggests that the Denver cell provided more favourable hydrodynamic conditions for particle–bubble collisions, a conclusion supported by the higher flotation rates of 0.65 min-¹ for the Denver cell compared to 0.44 min-¹ for the Leeds cell.
Moreover, both the fine (−25 µm) and intermediate (+25 –45 µm) size fractions required an additional 90 seconds to reach completion compared to the larger fractions. This finding is consistent with the conclusion of Pease et al. (2006) that finer particles generally require longer residence times in flotation processes than coarser particles.
The kinetic rate (k) was very low for the fine fraction (−25 µm) and increased with particle size, a trend consistent with the findings of Pyke et al. (2003) and Changuada et al. (2008). Particle size significantly affects the recovery rate, as highlighted by several researchers, including Trahar (1981), Rao (2004), and Jameson (2012). The maximum value of k was observed for the +45−75 µm fraction, but it decreased for the coarsest size fraction (+75 −106 µm). This behaviour can be attributed to the particle–bubble collection efficiency (Ek), which is determined by the product of collision efficiency (Ec), adhesion efficiency (Ea), and the preservation efficiency of the particle–bubble aggregate (Ep). Fine particles have low Ek due to poor Ec (i.e., limited particle–bubble collisions), while very coarse particles show reduced Ek because turbulence in the flotation cells tends to destroy the particle–bubble aggregates.
Pease et al. (2006) also demonstrated that finer particles require higher reagent dosages due to their larger specific surface area (cm²/g). To improve the recovery of fines to levels comparable to
Figure 8—Comparison of Denver (dotted lines) and Leeds (solid lines) cells for kinetics curves for flotation of the intermediate (+25−45 µm) and fine (−25 µm) size fractions of silica, with corresponding rate constant in the legend label
Figure 9—Effect of collector dosage on flotation recovery of fines (−25 µm) using the Denver (dotted lines) and Leeds (solid lines) cells, with corresponding rate constant in the legend label
those of the +25 µm size fractions, higher reagent dosages were investigated. The collector dosage was initially doubled, then further increased to 175 g/t. Figure 9 illustrates the impact of these increased dosages on fines recovery.
Despite doubling the dosage, recovery levels still fell short of those achieved for the +25 µm fractions. However, when the dosage was raised to 175 g/t of Flotigam EDA, a significant improvement in flotation response was observed, with recovery levels becoming comparable to those of the larger size fractions. Also, notably so, this required up to nine minutes to reach completion. As previously mentioned, finer particles typically require longer flotation times. This increase in reagent dosage enhanced overall flotation rates and potentially contributed to greater particle stabilisation, resulting in higher fines recovery from the froth.
Despite a significant increase in reagent dosage for the finest size fraction (−25 µm), the average k values remained relatively low at 0.34 min-¹ for the Denver cell and slightly lower at 0.23 min-¹ for the Leeds cell. This suggests that the Leeds impeller was less effective in promoting particle–bubble collisions compared to the Denver impeller, as all other conditions were kept constant. The observed increase in recovery rates can be attributed to the higher reagent dosage, which aligns with the findings of Pease et al. (2006).
Hydrodynamic
Figure 10 presents the results for the water-only system, showing a general trend of decreasing power number with increasing Reynolds number. This aligns with Westhuizen’s (2004) findings that the presence of air reduces the power number by decreasing the density effect around the impeller. The Leeds cell exhibited slightly higher turbulence intensities (higher Re) and power inputs than the Denver cell, except at 1000 rpm. Furthermore, increased turbulence
Comparison of Denver and Leeds laboratory flotation cells: Effect of particle size
intensity (higher NRe) and power input enhance bubble-particle collisions, especially for finer particles, leading to more efficient flotation, although excessive turbulence can potentially disrupt bubble-particle adhesion. These findings suggest that the Leeds cell is better suited for applications requiring higher turbulence intensity.
Figure 11 illustrates impeller performance in a slurry. The Leeds cell produced results similar to those in the water-only system, whereas the Denver cell maintained a stable power number of approximately 0.77 across the 1000 rpm–1500 rpm range. This stability highlights the Denver cell’s ability to generate sufficient turbulence while consuming less power.
In the absence of solids (Figure 10), both cells operated in the transitional region between laminar and turbulent flow, as evidenced by fluctuations in power numbers, suggesting the presence of large eddy vortices. However, in slurry conditions (Figure 11), the Denver impeller demonstrated a greater ability to overcome resistance from solid particles, with its power number rapidly approaching the fully turbulent regime. In contrast, the Leeds impeller did not exhibit this behaviour under the experimental conditions. These results indicate that the Denver cell produced more consistent fluid fields while operating with greater power efficiency.
Additionally, the Denver and Leeds cells had similar Froude numbers (2 < NFr< 4.5 and 2.1 < NFr < 4.7, respectively), indicating comparable solid suspension capabilities.
Conclusions
This study compared the performance of the Denver and Leeds laboratory flotation cells based on various particle size fractions and hydrodynamic characteristics to achieve comparable kinetic results. Hydrodynamic characteristics included dimensionless numbers such as power and Reynolds numbers, which revealed that these cells require different operating conditions in the presence and absence of particles. Leeds demonstrated higher power numbers under both conditions, indicating greater resistance to movement imposed by water and slurry due to a unique impeller-stator design. As a result, the Leeds impeller must dissipate more power to facilitate particle and bubble dispersion and promote particle–bubble collisions.
Consequently, the cells needed to be operated at different rotational speeds, 1200 rpm for the Denver cell and 1400 rpm for the Leeds cell, to achieve similar performance. The difference in operating speeds required to obtain similar results can be linked to hydrodynamic characteristics, which influence the manner of air
bubble creation in the cells and mineral recovery. These operating conditions were used for testing at narrow size fractions. The cells performed similarly at the coarser fractions, but there were some differences in the lower particle size range. This suggests that the impact of cell design, particularly impeller efficiency, becomes more pronounced at smaller particle sizes. The Denver cell reported higher recovery rates across all test ranges with faster kinetic rates.
The study emphasises the importance of cell design in flotation efficiency and highlights the critical role of particle size in flotation recovery. Fine particles require more reagents and longer flotation times (due to their larger surface area) to obtain optimum recoveries. Process efficiency will increase as a result of knowing how to optimise particle size distribution.
The results demonstrate that the Denver cell outperformed the Leeds cell, as the Leeds cell required a higher impeller speed to achieve similar recoveries. Overall, the Denver cell exhibited marginally superior performance, which can be attributed to its more versatile hydrodynamic characteristics, making it better suited for handling a broader range of particle sizes.
Acknowledgements
The authors wish to thank Anglo American for financial support and the University of Pretoria for use of the laboratory for this research. English editing of this manuscript was carried out by Prof. K. C. Sole.
References
Abo Dhaheer, M.S., Al-Rubaye, M.M., Alyhya, W.S., Karihaloo, B.L., Kulasegaram, S. 2015. Proportioning of self-compacting concrete mixes based on target plastic viscosity and compressive strength: part i - mix design procedure. Journal of Sustainable Cement-Based Materials, vol. 5, no. 4, 199–216. https://doi.org/10.1080/21650373.2015.1039625
Arbiter, N., Harris, C.C. 1962. Flotation machines. In: D.W. Fuerstenau (Editor). Froth Flotation 50th Anniversary Volume, Pap. 14. The American Institute of Mining, Metallurgical, and Petroleum Engineers, New York.
Changunda, K., Harris, M., Deglon, D.A. 2008. Investigating the effect of energy input on flotation kinetics in an oscillating grid flotation cell. Minerals Engineering, vol. 21, nos. 12–14, pp. 924–929. https://doi.org/10.1016/j.mineng.2008.03.015
El-Rahiem, F.H.A. 2014. Recent trends in flotation of fine particles. Journal of Mining World Express, vol. 3, pp. 63–79. https://doi.org/10.14355/mwe.2014.03.009
Figure 10—Power number vs. Reynolds number for water-air mixing in Denver and Leeds flotation cells
Figure 11—Power number vs. Reynolds number for quartz-water-air mixing in Denver and Leeds flotation cells between 1000 rpm and 1500 rpm
Comparison of Denver and Leeds laboratory flotation cells: Effect of particle size
Fuerstenau, M.C., Jameson, G.J., Yoon, R. H. 2007. Froth flotation: a century of innovation. Littleton, Colo.: Society for Mining, Metallurgy, and Exploration.
Harris, C.C. 1976. Flotation machines. In: Fuerstenau, M.C. Flotation: A.M. Gaudin Memorial Volume. AIME, New York, vol. 2, p. 788.
Harris, C.C. 1986. Flotation machines design, scale up and performance: data base. In: Somasundaran, P. (ed), Advances in mineral processing. Symposium honoring Nathaniel Arbiter on his 75th birthday, New Orleans. Littleton, p. 618–635.
Jameson, G.J. 2012. The effect of surface liberation and particle size on flotation rate constants. Minerals Engineering, 36-38, pp. 132–137. https://doi.org/10.1016/j.mineng.2012.03.011
Klimpel, R.R. 1984. Use of chemical reagents in flotation. Chemical Engineering, vol. 91, no. 18, pp. 75–79.
Leal Filho, L.S., Rodrigues, W.J., Ralston, J. 2002. Importance of hydrodynamics in coarse particle flotation. In: Ralston, J., Miller, J., Rubio, J. (eds). Flotation and flocculation: from fundamentals to applications. Snap Printing, Medindie (Australia), p. 203-212.
Libii, J.N. 2010. Using wind tunnel tests to study pressure distributions around a bluff body: the case of a circular cylinder.
Liddel, K.S., Dunne, R C. 1983. An evaluation of the Leeds laboratory-scale flotation cell. Council for Mineral Technology (Mintek), Randburg, South Africa.
Ma, Y., Li, S., Zhang, L., Li, H., Liu, Z. 2020. Numerical simulation on heat extraction performance of enhanced geothermal system under the different well layout. Energy Exploration & Exploitation, vol. 38, no. 1, pp. 274–297. https://doi.org/10.1177/0144598719880350
Murhula, E., Hashan, M., Otsuki, A. 2022. Effect of Solid Concentration and Particle Size on the Flotation Kinetics and Entrainment of Quartz and Hematite. Metals. vol. 13, no. 53, pp. 1–21. https://doi.org/10.3390/met13010053
Nelson, M.G., Lelinski, D. 2000. ‘Hydrodynamic Design of SelfAerating Flotation Machines’, Minerals Engineering, 13(10), 991–998. doi:10.1016/S0892- 6875(00)00085- 6
Newcombe, B., Bradshaw, D., Wightman, E. 2012. Development of a laboratory method to predict plant flash flotation performance. Minerals Engineering, vol. 39, pp. 228–238. https://doi.org/10.1016/j.mineng.2012.07.008
Nguyen, A.V. 2007. Flotation, Encyclopedia of Separation Science, pp. 1–27. https://doi.org/10.1016/B0-12-226770-2/00071-5
Norori-McCormac, A., Brito-Parada, P.R., Hadler, K., Cole, K., Cilliers, J.J. 2017. The effect of particle size distribution on froth stability in flotation. Separation and Purification Technology, vol. 184, pp. 240–247. https://doi.org/10.1016/j. seppur.2017.04.022
Pease, J.D., Curry, D.C., Young, M.F. 2006. Designing Flotation Circuits for High Fines Recovery. Minerals Engineering, vol. 19, no. 6, pp. 831–840. doi:10.1016/j.mineng.2005.09.056
Pyke, B., Fornasiero, D., Ralston, J. 2003. Bubble particle heterocoagulation under turbulent conditions. Journal of Colloid and Interface Science, vol. 265, no. 1, pp. 141–151. https://doi.org/10.1016/S0021-9797(03)00345-X
Ramlall, N.V. 2008. Development and modelling of a semi-batch. dissertation.
Rao, S.R. 2004. Flotation Kinetics and Fine Particle Flotation. In: Surface Chemistry of Froth Flotation. Springer, Boston, MA. https://doi.org/10.1007/978-1-4757-4302-9_14
Rodrigues, W.J., Leal Filho, L.S., Masini, E.A. 2001. Hydrodynamic Dimensionless Parameters and Their Influence on Flotation Performance of Coarse Particles. Minerals Engineering, vol. 14, no. 9, pp. 1047–1054. doi:10.1016/S08926875(01)00110-8
Ross, V. 2019. Key aspects of bench flotation as a geometallurgical characterization tool. Journal of the Southern African Institute of Mining and Metallurgy, vol. 119, no. 4, pp. 361–367. https://doi.org/10.17159/2411-9717/605/2019
Shabalala, N.Z.P., Harris, M., Leal Filho, L.S., Deglon, D.A. 2011. Effect of slurry rheology on gas dispersion in a pilot-scale mechanical flotation cell. Minerals Engineering, vol. 24, no. 13, pp. 1448–1453. https://doi.org/10.1016/j.mineng.2011.07.004
Safari, M., Harris, M., Deglon, D., Leal Filho, L., Testa, F. 2016. The effect of energy input on flotation kinetics. International Journal of Mineral Processing, vol. 156, pp. 108–115. https://doi.org/10.1016/j.minpro.2016.05.008
Smith, P.G., Warren, L.J. 1989. Entrainment of Particles into Flotation Froths. Mineral Processing and Extractive Metallurgy Review, vol. 5, nos. 1–4, pp. 123–145. https://doi.org/10.1080/08827508908952647
Souza Pinto, T.C., Braga, A.S., Leal Filho, L.S., Deglon, D.A. 2018. Analysis of key mixing parameters in industrial wemco mechanical flotation cells. Minerals Engineering, vol. 123, pp. 167–172. https://doi.org/10.1016/j.mineng.2018.03.046
Tabosa, E., Runge, K., Holtham, P., Duffy, K. 2016. Improving flotation energy efficiency by optimizing cell hydrodynamics. Minerals Engineering, vol. 96–97, pp. 194–202. https://doi.org/10.1016/j.mineng.2016.05.002
Trahar, W.J. 1981. A rational interpretation of the role of particle size in flotation. International Journal of Mineral Processing, vol. 8, no. 4, pp. 289–327. https://doi.org/10.1016/03017516(81)90019-3
Wang, L. 2016. Entrainment of fine particles in froth flotation, Thesis: Doctor of Philosophy, University of Queensland, Australia.
Wang, D., Liu, Q. 2021. Hydrodynamics of Froth Flotation and Its Effects on Fine and Ultrafine Mineral Particle Flotation: A Literature Review. Minerals Engineering, vol. 173. doi:10.1016/j.mineng.2021.107220
Westhuizen, A.P.P. 2004. The evaluation of solids suspension in a pilot scale mechanical flotation cell. Master dissertation. University of Cape Town, Cape Town, page 25.
Wills, B.A., Finch, J.A. 2015. Wills' Mineral Processing Technology: An Introduction to the Practical Aspects of Ore Treatment and Mineral Recovery. 8th edn., Oxford: Butterworth-Heinemann.
Yao, N., Liu, J., Sun, X., Liu, Y., Chen, S., Wang, G. 2021. A Rational Interpretation of the Role of Turbulence in Particle-Bubble Interactions. Minerals, vol. 11, no. 9, p. 1006. https://doi.org/10.3390/min11091006
Zheng, X., Johnson, N.W., Franzidis, J.-P. 2006. Modelling of Entrainment in Industrial Flotation Cells: Water Recovery and Degree of Entrainment. Minerals Engineering, vol. 19, no. 11, pp. 1191–1203. doi:10.1016/j.mineng.2005.11.005 u
GIW® MDX = BEST IN CLASS MILL CIRCUIT PUMP
Best In Class Pumps for Aggressive Slurries
Mining operators are seeing an increased demand for minerals and ores. Operators must now consider the costs of pump maintenance and replacement as well as the cost-benefit of utilizing fewer larger pumps versus several smaller pumps to perform the same work.
Mining operators are seeing an increased demand for minerals and ores. Operators must now consider the costs of pump maintenance and replacement as well as the cost -benefit of utilizing fewer larger pumps versus several smaller pumps to perform the same work.
The KSB GIW® MDX pump line delivers proven results in the harshest hard rock mining applications.
The KSB GIW® MDX pump line delivers proven results in the harshest hard rock mining applications. W: www.ksb.com/en-za T: +27-11-876-5600
W: www.ksb.com/en-za T: +27-11-876-5600
Affiliation:
Botswana International University of Science and Technology
Correspondence to: K. Fichani
Email: fichanik@biust.ac.bw
Dates:
Received: 20 Nov. 2023
Revised: 8 Apr. 2024
Accepted: 25 Mar. 2025
Published: May 2025
How to cite:
Fichani, K., Teseletso, L.S. Kaavera, J. Dintwe, T.K., Shemang, E., Matshediso, B.I. 2025. The impact of junior miners on the global supply of high-purity manganese sulfate monohydrate for the electric vehicle battery market. Journal of the Southern African Institute of Mining and Metallurgy, vol. 125, no. 5 pp. 233–242
DOI ID:
https://doi.org/10.17159/2411-9717/3196/2025
ORCiD: K. Fichani
http://orcid.org/0000-0001-9364-8217
LS. Teseletso
http://orcid.org/0000-0002-3984-4089
J. Kaavera
http://orcid.org/0009-0007-6878-5851
T.K. Dintwe
http://orcid.org/0000-0002-2739-624X
E. Shemang
http://orcid.org/0000-0003-2865-4148
B.I. Matshediso
http://orcid.org/0009-0001-2465-6114
The impact of junior miners on the global supply of high-purity manganese sulfate monohydrate for the electric vehicle battery market
by K. Fichani, L.S. Teseletso, J. Kaavera, T.K. Dintwe, E. Shemang, B.I. Matshediso
Abstract
The minerals sector continues to draw attention from policy makers who would want to see both old and new mineral projects add value to the minerals in the host country. The current technological development in the leaching of manganese oxide ores to produce a high-purity manganese sulphate monohydrate, a compound that is used in the manufacture of batteries for electric vehicles or long-life storage cells for the renewable energy market, has led to junior mining companies spearheading this technology in new projects across the globe. In this paper, we researched the high-purity manganese sulphate monohydrate project pipeline and used projected production volumes and C1 cash costs to construct the industry supply curve for six projects, including the K.Hill project owned by Giyani Metals Corporation near Kanye, in the southern part of Botswana. The aim was to determine the ideal conditions that would provide a comparative advantage for further local value addition to high-purity manganese sulphate monohydrate and to the end user product. The results showed that while the estimated C1 cash costs would place the K.Hill project as the second highest cost producer from among five projects, it would nonetheless have healthy profit margins due to the high projected price of high-purity manganese sulphate monohydrate. The study recommended policy options, which, if implemented, could further encourage the exploitation and value addition of battery metals in Botswana.
Keywords high-purity manganese sulfate monohydrate, battery metals, industry supply curve, electric vehicle
Introduction
The global supply of manganese, as estimated by the US Geological Survey for 202,1 was 20 million metric tonnes (Schnebele, 2022). A brief analysis of the US Geological Survey’s mineral commodity summary for manganese demonstrates the following: approximately 80% of manganese production during 2021 was from four countries, being South Africa (37%), Gabon (18%), Australia (16.7%), and China (6.5%). In terms of the distribution of known reserves, approximately 61% of these reserves occur in Africa, with South Africa accounting for the largest share, estimated at 42.7%, followed by Gabon and Ghana at 4.1% and 0.9%, respectively. Other countries with estimated significant shares of global manganese reserves are Australia and China at 18% and 4%, respectively. The characterisation and classification of manganese ores is based mainly on the manganese content, iron, and other impurities. Metallurgical ore contains more than 35% Mn, ferruginous ores and manganiferous ores contain 15% – 25% Mn and 5% – 10% Mn, respectively. Another common classification for manganese ore is high (>44% Mn), medium (30% – 44% Mn), and low (<30% Mn) (Ratshomo, 2013; Indian Bureau of Mines, 2014a). It is estimated that 90% of the global supply of manganese ore is used in ferrous and non-ferrous metallurgical applications for production of steel and ferroalloys (Indian Bureau of Mines, 2014b). In the ferrous metallurgical applications, approximately 30% of manganese ores are used in the steelmaking process itself as reactants in the de-oxidation and desulfurisation stages, with the remainder being used as ferroalloys in the steel product (Indian Bureau of Mines, 2014a). In steel making, the ferroalloys impart the requisite properties of strength and workability to the steel, a property for which no substitutes for manganese currently exist (Steenkamp et al., 2020). The non-ferrous metallurgical application is based on the production of high-purity electrolytic manganese metal (EMM) and electrolytic manganese dioxide. The former is used in the production of some alloys of non-ferrous metal like aluminium, copper, magnesium, and nickel (Van Zyl et al., 2016; Indian Bureau of Mines,
The impact of junior miners on the global supply of high-purity manganese sulfate
2014b). The main grades of alloys are high-carbon ferromanganese (HCFeMn: 65% – 80% Mn), medium-carbon ferromanganese (MCFeMn) and silicomanganese (SiMn: 50% – 74% Mn). The HCFeMn and SiMn require high-grade ores with over 44% Mn and low-grade ores at 33% – 35% Mn, respectively (Indian Bureau of Mines, 2014b). The remaining 10% of the global supply of manganese ore is used in non-metallurgical applications such as dry cell batteries, the glass industry, and a wide range of applications in the chemical and health industries (Van Zyl et al., 2016, Indian Bureau of Mines, 2014b).
In Botswana, the historical mining of manganese at Kgwakgwe Hill, which is currently referred to as K.Hill by the project owners, Giyani Metals Corporation, started in the pre-independence period and lasted fifteen years (from 1957 to 1972). The mining was focused on both the metallurgical and high-grade manganese oxide ores. Over the fifteen years of operation, the K.Hill area exported about 195 743 t of manganese ores (SRK Consulting UK Limited, 2020). In a Press Release dated April 6, 2023, Giyani Metals Corporation indicated that it had submitted an Environmental Impact Assessment (EIA) statement to the Department of Environmental Affairs of Botswana. The press release further states that the EIA statement, if approved, would lead to the company being granted a mining license for up to 25 years (Giyani Metals Corporation, 2023). The prospect of developing the K.Hill deposit into a manganese mine has generated interest from policy makers, as the government of Botswana issued an expression of interest in 2022, inviting possible joint venture partners for its electric mobility (e-mobility) programme under which the country intends to set up a factory for the manufacturing of electric vehicles (EV) (Kuhudzai, 2022). The K.Hill project would produce a precursor product, a high-purity manganese sulphate monohydrate (HPMSM), for the manufacturing of batteries for the EV market. The project will further provide some assurance about security of supply of the HPMSM to future local manufacturers of batteries for EVs. The other metals required in the lithium-ion batteries are nickel and cobalt. The Botswana Institute for Technology Research and Innovation (BITRI) and Process Research ORTECH (PRO) in Canada would carry out a study, which, if successful, would result in the setting up of a 30 000 t/a plant that would produce highgrade nickel and cobalt salts. This plant would be used to produce raw materials required to facilitate the production of EV and large energy storage batteries in Botswana (BITRI, 2022).
In this paper, we review the literature on publicly available information regarding the capacity and estimated unit costs to produce HPMSM by new or planned projects. This information is then used to construct the HPMSM industry supply curve to determine the likely competitiveness of the proposed K.Hill project. A further search of the literature is carried out to determine the ideal conditions that would provide a comparative advantage for further local value addition of the HPMSM beyond the precursor compound, manganese sulfate, and to the end user product, the battery or long-life storage cell for the EVs, and renewable energy markets. We conclude by providing policy options, which, if adopted, would further encourage the exploitation of battery metals in Botswana.
Geology, deposit size and mine production capacity
In this section, the major manganese mines in the world’s top three producer countries, i.e., South Africa, Australia, and Gabon, are compared to establish any similarities in ore types, size of resources, and manganese beneficiation processes as well as the end product at
the mine gate. Selected new manganese projects are also reviewed to determine if any influence exists on the extraction technology resulting from the expected growth in the future demand for manganese in the EV battery market.
In South Africa, manganese ore is mined from the Kalahari Manganese Field that is located in the Northern Cape. This field is a banded iron formation with inter-bedded units of manganese ore, and it contains over 90% of South Africa’s manganese resources. Globally, this field represents approximately 80% of land-based global resources of manganese ore. The field is estimated to host 4.2 Bt of manganese ore resources (Beukes et al., 2016). The overburden cover is shallow on the eastern side of the field and therefore amenable to exploitation by open-pit methods, such as at the Mamatwan mine, while it is deep seated on the western side, up to 1400 m. The depths of underground workings are estimated at approximately 400 m at the Wessels and Nchwaning mines (Beukes et al., 2016). The Kalahari Manganese Field covers an area that is approximately 15 km by 35 km in the EW and NS directions, respectively. There are two major ore types present. The first is a low-grade primary sedimentary-type ore rich in carbonates, primarily calcites, and dolomite with braunite (2Mn2O3MnSiO3; 64.3% Mn) as the main manganese-bearing mineral. The second ore type is a high grade, structurally-controlled hydrothermal ore comprised mostly of oxides, mainly braunite and braunite II (Ca(MnFe)14SO24), some hausmannite (Mn3O4; 72% Mn), bixbyite ((Mn,Fe)2O3), and hematite (Fe2O3) which is found in the northern part of the deposit (Tsikos et al., 2003; Ratshomo, 2013).
Manganese ore production in Australia is mainly concentrated in the Groote Eylandt deposits, Gulf of Carpentaria in the Northern Territory, and in the Woodie Woodie mine, eastern Pilbara Craton in Western Australia (NSW Department of Primary Industries, nd). The geology of Groote Eylandt is described as Early to Middle Cretaceous Mullaman beds (sandstone, claystone, pebble gravel, and manganese marl), unconformably overlain by lateritic Tertiary conglomerate. At the Groote Eylandt mine, the mineralised zones average 3 m in thickness, approximately 150 km2 in areal extent, and the overburden cover ranges from 3 m – 12 m. The Groote Eylandt deposit is thought to have formed as a shallow water marine sedimentary deposit on a Proterozoic basement with postenrichment processes. Conversely, the Woodie Woodie mine ores are thought to have formed from both supergene enrichment of manganiferous sedimentary rocks and filling of cavities and fissures in dolomite and fault zones. The ores consist mainly of pyrolusite (MnO2; 60% – 63% Mn), cryptomelane (KMn8O16; 59.8% Mn), and minor manganite (Mn2O3H2O; 62.5% Mn), and trace amounts of other manganese minerals (NSW Department of Primary Industries, nd).
The production of manganese ores is also widespread in West Africa, with Gabon hosting the most significant resources outside South Africa. The manganese deposits in Gabon are supergene manganocrete that developed on relatively unmetamorphosed and undeformed manganese-bearing black carbonaceous shale of the Francevillian Supergroup that overlies the Archaean basement in the northwestern part of the Congo Craton. The deposits are large and shallow and therefore exploited by open pit mining. The manganese resource is estimated to be 325 Mt at 49% Mn in the washed product, with a regional potential in the order of 100 Mt – 200 Mt. The ore minerals in the Franceville area consists of cryptomelane and pyrolusite with some nsutite and lithiophorite set in a goethitic, clay, and quartz matrix (Beukes et al., 2016).
The impact of junior miners on the global supply of high-purity manganese sulfate
The preceding three descriptions are examples of large-scale manganese deposits that are exploited by companies based in the world top three manganese ore-producing countries. Jupiter Mines, which is a 49.9% shareholder in the Tshipi é Ntle Manganese Mining Proprietary Limited (Tshipi) located in the Kalahari Manganese Field in the Northern Cape, RSA, is already a major manganese ore producer with target production of 3.6 Mt of manganese ore of which 600 kt is low grade (30% Mn). The low-grade ore would be supplied to a proposed HPMSM facility to be located in either Canada, USA, or Europe (Jupiter Mines Ltd, 2024). It joins a list of five other HPMSM projects at different project study stages, such as the Giyani Metals Corporation’s K.Hill project in Botswana (Giyani Metals Corporation, 2022; Keating, 2023), Canadian Manganese’s Woodstock manganese project in New Brunswick, Canada (Sprott Capital Partners Equity Research, 2022), Manganese X Energy Corporation’s Battery Hill project near Woodstock, New Brunswick (Manganese X Energy Corp., 2022), Euro Manganese’s Chvaletice Manganese project, Chvaletice, Czech Republic (Tetra Tech Canada Inc., 2022), and the Butcherbird project by Element 25 in Western Australia, of which the project will produce manganese concentrates to be sold under an off-take agreement to the company’s HPMSM facility in Louisiana (Element 25 Limited, 2022). The common business objective of these projects is to produce HPMSM for the expected high demand in the EV battery market.
Manganese extraction technologies
One of the key factors influencing the choice of extraction
technology for manganese between pyrometallurgical approaches employing reduction smelting in either blast furnace or submerged arc electric smelting furnaces (SAF), or hydrometallurgy involving leaching followed by solvent extraction and electrowinning, is the size of the deposit. Table 1 indicates that, for the major manganese ore producers, there is very limited production of ferroalloys in their home countries apart from China, which is the world’s leading producer of steel and manganese ferroalloys such as SiMn and HCFeMn. It is therefore apparent that the manganese value chain for large or world class deposits follows the pyrometallurgical route as depicted in Figure 1.
Source: Compiled by the authors based on (Steenkamp et al., 2018; Van Zyl, et al., 2016)
Figure 1—A simplified value chain for manganese for world class deposits
Recent ore reserves or resources and production from selected major manganese ore producing countries
Mine, company and country
Mamatwan, Hotazel Manganese Mines, Northern Cape, South 32 Limited, RSA.1
Tshipi é Ntle Manganese Mining Pty Ltd (Tshipi), Jupiter Mines Ltd, Northern Cape, RSA.2
Groote Eylandt Mining Company, South 32 Limited Northern Territory, Australia.1
Moanda mine, ERAMET (Comilog), Gabon.3
Source:
As of June 2022: Proven and probable-45Mt @ 36.2% Mn; Total resources-72 Mt @ 35% Mn; LoM as at Jan 2022-14 years.
As of February 2022: Measured –129.21 Mt @33.85% Mn; Indicated – 70.15 Mt @ 33.28% manganese and inferred – 225.80 Mt @ 32.77% Mn. Total mineral resources – 425 Mt @ 33.18% Mn.
As of June 2022: Proven and probable- 35 Mt @ 43% Mn and 4.8 Mt Sands @ 40% Mn; Total resourses- 138 Mt @ 43% Mn and 9.0 Mt sands @ 19.5% Mn; LoM as at Jan 2022-3.9 years.
Company webpage attributes 25% of global reserves and grades at 30% Mn.2
1Compiled by the authors from (South32 Limited, 2022)
Ore production in 2022 was 2.069 Mt.1
Ore production capacity is 3.6 Mtpa. In 2023, it was 3.34 Mt down from 3.68 Mt in 2022.
Washed lump ore, and Sinter for SAF smelting locally and for export.
Washed lump ore and future HPMSM product for EV batteries.
Ore production in 2022 was 3.363 Mt.1
Washed lump ore and fines all exported.
Estimated ore production is 4.0 Mt but it is projected to increase to 7.0 Mtpa.2
2Compiled by the authors from (Jupiter Mines Ltd, 2024; Jupiter Mines Ltd, 2023)
3Compiled by the authors from (Comilog, n.d.)
Lump ore, sinter and ferroalloys (silico manganese) for the export market.
Table 1
The impact of junior miners on the global supply of high-purity manganese sulfate
The manganese project pipeline is presented in Table 2. It is apparent that these projects are at different stages of study, ranging from scoping studies or preliminary economic assessment to feasibility studies. The K.Hill project in Botswana is at the feasibility study stage but due to a recent fourfold increase in its Indicated Mineral Resources from 2.0 Mt to 8.6 Mt, the company is planning to conduct a preliminary economic assessment to consider this improvement in resources and possibly increasing the life of mine to beyond 25 years (Keating, 2023). The Butcherbird project, has completed a feasibility study to produce HPMSM from manganese
concentrates. For the five other projects, which consist of two projects in Canada, one each in the Czech Republic and Botswana, it will produce HPMSM. Lastly, the HPMSM project by Jupiter Mines Ltd is at the scoping study stage and its HPMSM facility may be located in either Canada, the US, or Europe.
The processing methods of the ore would depend on the ore type with direct leaching of ores or concentrates using dilute sulfuric acid applied in manganese carbonate ores while reduction leaching is applied in oxide ores (Manganese X Energy Corp, 2021; Tetra Tech Canada Inc, 2019). If the ores are manganese oxides,
Preliminary and advanced stage manganese projects for producing HPMSM for the EV battery market
Project, company and country Manganese ore minerals Resources / reserves Ore
Battery Hill, Manganese X Energy Corp., Canada1
Butcherbird HPMSM project, Element 25, Australia2
Manganese carbonate and manganese carbonatesilicate-oxide.
Manganiferous shale (Pyrolusite).
Chevaltice, Euro Manganese, Czech Republic3
K.Hill, Giyani Metals Corp., Botswana4
Tshipi é Ntle Manganese Mining Pty Ltd (Tshipi), Jupiter Mines Ltd, Northern Cape, RSA5
Tailings from old pyrite mines; 81% of ore is manganese carbonate and the rest manganese carbonate-silicates.
Manganiferous shale, that occurs in layers 3–4 m thick and are high grade manganese oxides, Pyrolusite (MnO2)
Low grade - primary sedimentary-type ore rich in carbonates High grade – oxides, mainly braunite and braunite II
Woodstock project, Canadian Manganese Company Inc., New Brunswick, Canada6
Manganese carbonates, mainly Rhodochrosite.
Source: Compiled by the authors from:
1(Manganese X Energy Corp., 2022)
2(Element 25 Limited, 2022)
3(Tetra Tech Canada Inc., 2019)
Measured –11.26 Mt @6.75% Mn; Indicated – 23.60 Mt @ 6.26% Mn and Inferred – 25.91 Mt @6.66% Mn.
Measured – 16.0 Mt @11.6% Mn; Indicated – 41Mt @10% Mn, and Inferred – 206 Mt @ 9.8%.
Indicated – 464 kt @7.85% Mn; Measured – 26.496 Mt @7.32% Mn.
Indicated Resource of 8.6 Mt ore at 15.2 MnO at a cut-off grade of 7.3% MnO. Inferred Resource of 6.1 Mt at 14.1% MnO.
As at February 2022: Measured – 129.21 Mt @33.85% Mn; Indicated – 70.15 Mt @ 33.28% Manganese and Inferred –225.80 Mt @ 32.77% Mn. Total mineral resources – 425 Mt @ 33.18% Mn.
Direct leaching of RoM ore at 365kt per annum to produce 68 kt/a HPMSM in year 3 from project start date.
1.3 Mtpa RoM beneficiated to produce lump ore fines concentrates; HPMSM future project based on concentration feed.
Direct leaching of RoM ore at 1,1 Mtpa starting in year 4 from project start date to produce 120 -150 kt / year a HPMSM.
Direct leaching of RoM ore at 200 kt per year starting in 2025 to produce 120 kt/ year of HPMSM.
Low grade ore production of 600 kta.
HPMSM plant located in either Canada, USA or Europe.
Estimated HPMSM capacity is 80 – 120 kt/year in 2028.
Inferred – 44.8 Mt @9.9% Mn. Initially 700tpd RoM and direct leaching of concentrate to produce 60kt/a of HPMSM. Potential for 150kt/a by 2040.
4(Keating, 2023; Giyani Metals Corporation, 2022; SRK Consulting UK Limited, 2020)
5(Jupiter Mines Ltd, 2024; Jupiter Mines Ltd, 2023)
6(Sprott Capital Partners Equity Research, 2022)
Recovery – 78%; HPMSM.
Lump ore, and In future HPMSM and HPEMM.
Recovery – 59.2% (Overall) HPEMM HPMSM.
Recovery – 88.5%; HPMSM.
HPMSM product for EV batteries.
Recovery – 77.1%; HPMSM.
Table 2
The impact of junior miners on the global supply of high-purity manganese sulfate
studies have demonstrated that reduction leaching is possible using a variety of reductants such as sulfur dioxide, sucrose, ferrous sulphates, and many others (Giyani Metals Corporation, 2022; Kumar, Purcell, 2019). Another pathway that manganese producers are likely to adopt is to extend the EMM pathway by leaching this product, which is about 99.7% Mn, followed by evaporation and crystallisation to produce HPMSM at 99.9%Mn (Winjobi, Kelly, 2021). The Chevaltice process flow diagramme demonstrates this (Tetra Tech Canada Inc., 2022). Figure 2 depicts a block diagramme for the production of HPMSM from the direct leaching of manganese carbonate ores and reductive leaching of manganese oxide ores.
Structure of the EV market
At the global level, the EV market can be divided geographically into four major regions of Europe, China, US, and the rest of the world as depicted in Table 3, and in general, robust growth in the production of EVs is projected in all the four world regions. This may be attributed to a number of factors, which include meeting the consumer concerns on issues such as: driving distance, availability of charging points, price premium over the ICE vehicles,
Table 3
The EV markets and the top five selling brands Region / auto maker1
Europe
1. BMW
2. VW
3. Renault
4. Tesla
5. Mercedes
China
1. BYD
2. BJEV
3. Roewe
4. Tesla
5. Chery
US
1. Tesla
2. Chevrolet
3. Nissan
4. Toyota
5. Ford
Rest of World
1. Nissan
2. Tesla
3. Mitsubishi
4. Toyota
5. Hyundai
Sources: Compiled by the authors from the following sources:
1(Li, et al.,2021)
2(International Energy Agency, 2022)
3(Chege, 2021)
4(MacIntosh, et al., 2022)
states of California, New York and Massachusetts - 100% by 20354
by 20504
Figure 2—A simplified block diagram for the production of HPMSM for EV batteries
Source: Compiled by the authors based on (Giyani Metals Corporation, 2022; Manganese X Energy Corp, 2021; Sprott Capital Partners Equity Research, 2022)
The impact of junior miners on the global supply of high-purity manganese sulfate
introduction of laws that penalise the continued use of ICE vehicles through emission taxes and prohibition of access to cities, and providing incentives for switching to EVs (Deloitte, n.d.). Fiscal incentives such as price subsidies and, to a lesser extent, recurrent fees paid by vehicle owners such as emissions taxes, coupled with a well-developed, accessible, and harmonised payment system and a high-quality charging network, are the main drivers that influence the uptake of EVs (MacIntosh et al., 2022; Vanpée et al., 2022). Disregarding the policy interventions, there is a general agreement among analysts that the cost of the battery pack for the EVs must reduce from USD121/kWh in 2021 to below USD100/kWh (MacIntosh et al., 2022).
The top five global auto maker’s as indicated in Table 3 accounted for some 41% of global car sales in 2020 compared to the top five global battery makers who accounted for 83% of the sales in the same year. This reflects an oligopolistic market structure in EV trade, which may lead to higher battery prices. This situation may lead to the top five auto makers entering into joint ventures with battery makers to achieve a lower price for the EV batteries (Goldman Sachs Equity Research, 2022).
Supply outlook of HPMSM for the EV battery market
Global production of HPMSM is dominated by China, which in 2017 produced some 80 kt, representing 87% of global production, with the balance of about 12 kt or 13% attributed to a US company with production plants in both Mexico and Belgium (Tetra Tech Canada Inc., 2019). It was estimated that China was planning to increase its HPMSM capacity to 940 kt/a by 2022. An HPMSM plant with a capacity of some 30 kt/a came into production in Indonesia (Tetra Tech Canada Inc., 2019). The planned HPMSM projects are looking at the regional demand for HPMSM in their regions, for instance both the Woodstock and Battery Hill projects by Canadian Manganese Company Inc. and Manganese X Inc. are targeting mainly the North American EV battery market (Manganese X Energy Corp., 2022) while the Chevaltice project by Euro Manganese is targeting the European EV battery market (Tetra Tech Canada Inc., 2019). If all the six HPMSM projects in Table 2 are successfully developed and brought into production over the near term, then they would add approximately 530 kt HPMSM to the global market for HPMSM.
Demand outlook for HPMSM for the EV battery market
As approximately 90% of global manganese production is used in the steel industry, and given that there is no substitute for manganese in steel making, it justifies the assumption that the growth in global manganese consumption will follow that for steel. Global steel production is forecast to grow from some 1.8 Bt in 2022 to 2.3 Bt in 2030, which represents a compound annual growth rate of 3% over the period (Globenews Wire, 2023). The demand for EVs, and therefore EV batteries, is driven by commitments that member countries made at the UN Conference of the Parties, COP26, that was held in Glasgow in October 2021. The member countries agreed to timelines for adopting zero-emission vehicles (ZEV) with countries setting target dates over the period 2025 to 2050 but with the majority of signatories, which excluded developed countries with leading automobile makers such as the U.S., Germany, Japan, South Korea, and China, setting targets for the period 2025 to 2040 (MacIntosh et al., 2022). As part of this policy, there has been tremendous growth in EVs and this is forecast to continue at a compound annual growth rate of 30%, from a base of 21 million vehicles in 2022, excluding two and three wheelers, to 200 million vehicles by the year 2030 (International Energy
Agency, 2022). The exponential growth in EV demand is forecast to create a demand for lithium-ion batteries that will, in turn, lead to an exponential growth in demand for HPMSM. Some forecasts indicate that HPMSM demand will grow from 225 kt in 2022 to over 600 kt by 2030 (Sprott Capital Partners Equity Research, 2022).
Materials and methods
Production theory defines the industry supply or cost curve as a horizontal summation of the individual firm’s supply curves. Direct cash production costs (C1) are defined to include royalties, costs of mining, processing, and site general and administration costs less any byproduct credits. The C1 cash costs for the six HPMSM projects are presented in Table 4. Half of these projects are at the scoping study stage, represented as Class 5, under the American Association of Cost Engineers (AACE) while the other half is at the feasibility study stage, represented as Class 3. The level of accuracy reported is for the particular study and not necessarily the generic class.
The six HPMSM projects are arranged in ascending order by their C1 cash costs expressed in USD/t of HPMSM produced, as shown in the last column of Table 5. A MicroSoft Excel based algorithm, which enables the computation to transform the data into a form for plotting the cost curve for the six projects was then used (Tholana et al., 2013). The data for plotting the HPMSM supply curve are presented in Table 6.
Results and discussion
The HPMSM supply curve for the six projects is depicted in Figure 3, which can be used to analyse the competitiveness of these projects relative to one another. The supply curve ranks the Chevaltice Manganese Project as the highest cost producer with direct cash costs at USD2,319/t, followed by K.Hill at USD1,846/t and Butcherbird at USD1,188/t HPMSM produced. These three projects are at the Class 3 or feasibility study stage with an accuracy level in the range -15% to +20% as shown in Table 4. The other three projects were at the scoping study stage or Class 5 with their level of accuracy also shown in Table 4 being in the range of -30% to +50%. The lowest cost project would be Battery Hill at USD654/t followed by Woodstock at USD675/t and Jupiter Mines at USD1,020/t of HPMSM produced. When the different levels of accuracy of the cost estimates in Table 4 are taken into account, the ranking of these projects on the computed supply curve remains unchanged. All projects were modelled on different HPMSM price assumptions and they demonstrated positive margins and real net present values (NPV). For instance, the Battery Hill project’s preliminary economic assessment assumed a long term price of USD2,900/t HPMSM while the K.Hill project used a long term price of USD2,993/t HPMSM in 2026 rising to USD3,918 in 2030 and remaining at that level to end of mine life in 2035 even though their sales are to the same markets in the US and Western Europe as would the Battery Hill project.
Apart from the K.Hill project, which would be mining and processing a manganese oxide ore, the other five projects presented in Table 4 are based on manganese carbonate ores. The oxide ore would require reduction leaching using dilute sulfuric acid with sulfur dioxide as a reductant, while the other five projects would apply direct leaching of the manganese carbonate ores using dilute sulfuric acid only. The high cost for the Chevaltice Manganese Project, at USD2,106/t HPMSM, sits at the high end of this curve. The high production costs for Chevaltice can be explained by the processing route in which the project first produces high-purity electrolytic manganese metal (HPEMM) followed by the production
The impact of junior miners on the global supply of high-purity manganese sulfate
Table 4
C1 Cash costs and their level of accuracy
Source: Cost data are compiled by the authors from the following sources:
1Table 21-3 Total Operating costs over LOM, Technical Report on the Preliminary Economic Assessment of the Battery Hill Manganese Project Woodstock, New Brunswick, Canada (Manganese X Energy Corp., 2022)
2Feasibility study - battery grade high purity manganese processing facility (Element 25 Limited, 2023)
3(Sprott Capital Partners Equity Research, 2022)
4Table 21-3 LOM Average operating cost summary, Technical Report and Feasibility Study for the Chvaletice Manganese Project, Czech Republic (Tetra Tech Canada Inc., 2022)
5Operating costs, pp 7 (Jupiter Mines Ltd. 2024)
6Table 4 Summary of Project operating expenditure per tonne processed (Giyani Metals Corporation, 2022)
Table 5
Input data for computing C1 cash costs in USD/t milled
Notes:
aOn average the Chevaltice project will produce 47.5 kt of HPEMM /year and 67% of this would be converted to HPMSM yielding 98.6 kt/a (Tetra Tech Canada Inc., 2022)
bG&A – General and administrative costs
Source: Cost data are compiled by the authors from the following sources:
1Table 21-3 Total Operating costs over LOM, Technical Report on the Preliminary Economic Assessment of the Battery Hill Manganese Project Woodstock, New Brunswick, Canada (Manganese X Energy Corp., 2022)
2Feasibility study - battery grade high purity manganese processing facility (Element 25 Limited, 2023)
3Operating costs, pp7 (Jupiter Mines Ltd. 2024);
4(Sprott Capital Partners Equity Research, 2022);
5Table 4 Summary of Project operating expenditure per tonne processed (Giyani Metals Corporation, 2022)
6Table 21-3 LOM Average operating cost summary, Technical Report and Feasibility Study for the Chvaletice Manganese Project, Czech Republic (Tetra Tech Canada Inc., 2022)
The impact of junior miners on the global supply of high-purity manganese sulfate
Table 6
Data for plotting the HPMSM cost curve
Source: HPMSM production data from sources cited for Table 5 while plotting data were computed by the authors applying the cost curve algorithm (Tholana et al., 2013)
of HPMSM from the HPEMM. The costs of these processing stages are estimated to constitute 33% and 13%, respectively, of the on mine direct production costs (Tetra Tech Canada Inc., 2022). It is estimated that approximately two thirds of the HPEMM produced from the electrowinning stage is further dissolved followed by purification and crystallisation to produce the HPMSM product (Tetra Tech Canada Inc., 2022).
The processing costs per tonne of ore milled are presented in Table 5 and they demonstrate that the K.Hill project, at USD636/t of ore milled would be approximately four times greater than those for the Battery Hill project at USD153/t of ore milled. The high processing costs for K.Hill are attributed to international import freight cost of reagents, which are estimated at 36% of the total reagent costs for the project. As expected, for the mining stage, costs would be lowest for Chevaltice as it would be treating tailings and therefore would not require drilling and blasting for excavating the manganese ore.
Possible policy options for attracting EV and EV battery manufacturers
The government of Botswana has made its intentions known regarding its e-mobility programme and by taking action through an expression of interest for a joint venture partnership with the private sector in developing EVs (Kuhudzai, 2022). Of the base
metals that would be used as inputs in electric cars, the country currently produces only copper, but plans by a local subsidiary of Premium Nickel Resources Limited (PNRL) to revive the BCL Ltd mines, which would produce nickel, copper, and cobalt, would lead to a re-start of the mining operations by 2026 (Mguni, 2023).
The e-mobility programme aims to encourage the setting up of production facilities for EVs and EV batteries. Depending on the structure that would be adopted by EV investors, they may well choose to integrate and produce their own EV batteries. The question then is what policy options are there for are supporting an integrated EV and EV battery maker.
Many researchers have studied the types of policies and incentives in the world’s leading markets for EVs. These are categorised based on their aim, which may be to encourage diffusion or dissemination of products through support to the supply and demand sides of the EV market; to encourage research and development (R&D) for the development of technologies to support the growth of e-mobility, and finally encourage the development of infrastructure such as charging infrastructure for the EV market (see Sousa, Costa, 2022 for a review). The policy approach to encourage diffusion of EVs in the leading EV market countries is fiscal incentives in the form of price subsidies and lower licensing costs. These subsidies are paid out of the fees paid by owners of the ICE vehicles (Kohn et al., 2022).
It is apparent that the major EV market countries have set targets by which they plan to have reached zero emissions of greenhouse gases (MacIntosh et al., 2022). The policies to achieve these targets rely on the adoption of zero emission vehicles in these countries. As the EV market is non-existent in Botswana and the Southern African region, this suggests that the policy options would have to include all of the three categories to first conduct R&D in the use of the available raw materials in the production of EVs and EV batteries, the provision of a robust charging network infrastructure, and lastly to determine the fiscal incentives to cushion the consumer in the purchase of EVs.
From a raw materials perspective, the availability of nickel, copper, cobalt, manganese, and HPMSM would provide comfort over the security of supply to an EV or EV battery manufacturer located in the country.
Conclusions
This study set out to review publicly available information on
Figure 3—The HPMSM cost curve for the five selected projects
Source: Compiled by the authors based on (Giyani Metals Corporation, 2022; Manganese X Energy Corp, 2021; Sprott Capital Partners Equity Research, 2022)
The impact of junior miners on the global supply of high-purity manganese sulfate
planned manganese projects globally that would produce HPMSM for use in EV batteries because of the policy-induced growth in the global demand for EVs. The approach adopted involved gathering both the C1 project cash costs and the estimated production capacities for each of HPMSM projects studied and plotting a cost curve for the estimated supply of HPMSM from the six projects in this study. This cost curve was then used to assess the cost competitiveness of the K.Hill project, which is currently at an advanced stage of exploration by Giyani Metals Corporation. We conclude that the K.Hill project would be the second highest cost producer of HPMSM and this is largely due to the processing costs for reduction leaching of the oxide ore with sulfuric acid and using sulfur dioxide as a reductant. While its costs, at USD1,846/t HPMSM are lower than those for the Chevaltice Manganese Project at USD2,319/t HPMSM, they would still cost about twice as much as each of the three projects on the lower end of the cost curve. The three projects at the lower end of the HPMSM supply curve are based on manganese carbonate ores that are amenable to direct leaching using only dilute sulfuric acid. It is worth noting, however, that projected long term prices for the HPMSM still present a substantial margin over the C1 cash costs for the K.Hill project. Regarding the policy option to encourage the development of the EV and EV battery industry in the country, the study concludes that the local availability of raw materials would offer security of supply to any future investor.
References
Beukes, N.J., Swindell, E.P.W., Wabo, H. 2016. Manganese deposits of Africa. Episodes, vol. 39, pp. 285–317.
BITRI. 2022. The Bitri-Pro High Grade Salts Extraction Project Background and Goals. https://www.bitri.co.bw/wp-content/ uploads/2022/10/BITRISept2022.pdf [Accessed January 27, 2023].
Chege, W.W. 2021. Electric Vehicles-Market Intelligence Report. https://www.westerncape.gov.za/110green/sites/green. westerncape.gov.za/files/atoms/files/EV_MIR_29_3_22_FINAL. pdf [Accessed February 25, 2023].
COMILOG. n.d. World’s second largest producer of high-grade manganese ore. https://comilog.eramet.com/en/comilog/ comilog-at-a-glance/ [Accessed June 24, 2023].
Deloitte. n.d. Electric vehicles - Setting a course for 2030. https://www2.deloitte.com/content/dam/insights/us/ articles/22869-electric-vehicles/DI_Electric-Vehicles.pdf [Accessed February 25, 2023].
Element 25 Limited. 2023. Feasibility study - battery grade high purity manganese processing facility. https://www.element25. com.au/site/pdf/d53d05d9-3bc8-4da9-b671-56d960c3bb28/ HPMSM-Feasibility-Study-Delivers-Outstanding-Economics. pdf [Accessed June 26, 2023].
Firebird Metals Limited. 2022. Impressive results from manganese concentrate scoping study at Oakover, Subiaco, Western Australia. https://minedocs.com/22/Oakover-PEA-7202022.pdf [Accessed February 24, 2023].
Giyani Metals Corporation. 2022. Giyani Announces the Results of its Feasibility Study for the K.Hill Battery-Grade Manganese Project. https://www.globenewswire.com/en/ news-release/2022/11/14/2554778/0/en/Giyani-Announces-theResults-of-its-Feasibility-Study-for-the-K-Hill-Battery-GradeManganese-Project.html [Accessed February 8, 2023].
Giyani Metals Corporation. 2023. Giyani announces submission of Environmental Impact Statement for K.Hill. https://www. globenewswire.com/news-release/2023/04/06/2642527/0/en/ Giyani-Announces-Submission-of-Environmental-ImpactStatement-for-K-Hill.html [Accessed May 18, 2023].
Globenews Wire. 2023. Global Steel Market to Reach 2.3 Billion Metric Tons by 2030. https://www.globenewswire.com/ news-release/2023/02/16/2609919/0/en/Global-Steel-Marketto-Reach-2-3-Billion-Metric-Tons-by-2030.html [Accessed February 22, 2023].
Goldman Sachs Equity Research. 2022. Electric Vehicles: What’s Next VII: Confronting Greenflation. https://www.goldmansachs. com/insights/pages/gs-research/electric-vehicles-whats-nextvii-confronting-greenflation/report.pdf [Accessed February 26, 2023].
Indian Bureau of Mines. 2014a. Manganese ore-vision 2020 and beyond. https://ibm.gov.in/writereaddata/ files/11052014103838Manganese_Ore_Vision_2020_and_ Beyond.pdf [Accessed January 27, 2023].
Indian Bureau of Mines. 2014b. Market survey on manganese ore. http://ibm.nic.in/writereaddata/ files/07142014124825Market%20Survey_Manganese%20Ore. pdf [Accessed January 31st, 2023].
International Energy Agency. 2022. Global EV Outlook 2022-Securing supplies for an electric future. www.iea.org [Accessed February 22, 2023].
Jupiter Mines Ltd. 2024. Scoping study: High purity manganese sulphate monohydrate (HPMSM) project. https://www. jupitermines.com/cproot/1197/3/Scoping-Study-High-PurityManganese-Sulphate-Project.pdf [Accessed May 24, 2024].
Jupiter Mines Ltd. 2023. Annual Report, 2023. https://www. jupitermines.com/cproot/1155/3/20230531%20Jupiter%20 Mines%20Annual%20Report%202023%20(Final).pdf [Accessed May 24, 2024].
Keating, D. 2023. Giyani Metals Announces 310% Increase in Indicated Mineral Resources for K.Hill Project. https:// www.juniorminingnetwork.com/junior-miner-news/pressreleases/1007-tsx-venture/emm/144877-giyani-announces-310increase-in-indicated-mineral-resources-for-k-hill-project.html [Accessed July 19,2023].
Kohn, E., Huang, C., Kong, N., Hardman, S. 2022. Electric Vehicle Incentives in 15 Leading Electric Vehicle Markets. http://www.phev.ucdavis.edu [Accessed March 5th, 2023]
Kuhudzai, R.J. 2022. Botswana invites local and international companies to express an interest in manufacturing EVs in the country. https://cleantechnica.com/2022/04/15/botswanainvites-local-international-companies-to-express-an-interest-inmanufacturing-evs-in-the-country/ [Accessed 27/1/2023].
Kumar, S.M., Purcell, W. 2019. Reducing agents in the leaching of manganese ores: A comprehensive review. Hydrometallurgy, vol. 187, pp. 168–186.
Li, S., Wang, B., Yang, M., Zhang, F. 2021. The global diffusion of electric vehicles-lessons from the first decade. Policy research working paper: 8892 https://documents1.worldbank.org/ curated/en/225111639490843204/pdf/The-Global-Diffusion-ofElectric-Vehicles-Lessons-from-the-First-Decade.pdf [Accessed February 25, 2023].
Macintosh, R., Tolomiczenko, S., Horn, G.V. 2022. Electric Vehicle Market Update-Manufacturer Commitments & Public Policy Initiatives Supporting Electric Mobility in the U.S. & Worldwide. https://www.sustainability.com [Accessed February 25th, 2023].
Manganese X Energy Corp. 2021. Battery Hill Project Mineral Resource Estimate. https://www.manganesexenergycorp.com/ wp-content/uploads/2021/08/MNX-43-101-Battery-HillAug-2021.pdf [Accessed February 20, 2023].
Manganese X Energy Corp. 2022. Manganese X Energy Announces Positive PEA for its Battery Hill Project. https://www. manganesexenergycorp.com/manganese-x-announces-positivepea-for-its-battery-hill-project/
The impact of junior miners on the global supply of high-purity manganese sulfate
Mguni, M. 2023. PNR expects 2026 start date for BCL [Online]. Gaborone. Available: https://www.mmegi.bw/business/pnrexpects-2026-start-date-for-bcl/news [Accessed November 1, 2023].
NSW Department of Primary Industries. 2022. Manganese https:// meg.resourcesregulator.nsw.gov.au/sites/default/files/2022-11/ manganese.pdf
Ratshomo, K. 2013. South Africa's manganese industry developments 2004-2011. https://www.dmr.gov.za/LinkClick. aspx?fileticket=g2xSDGAbOug%3D&portalid=0. [Accessed February 1, 2023]
Schnebele, K.E. 2022. Mineral Commodity Summaries –Manganese. https://pubs.usgs.gov/periodicals/mcs2022/ mcs2022-manganese.pdf [Accessed May 19, 2023].
Sousa, C., Costa, E. 2022. Types of Policies for the Joint Diffusion of Electric Vehicles with Renewable Energies and Their Use Worldwide. https://doi.org/10.3390/en15207585
South32 Limited. 2022. Annual Report. https://www.south32.net [Accessed June 24,2023].
Sprott Capital Partners Equity Research. 2022. Canadian Manganese-initiating coverage-best exposure to manganese for battery markets. https://sprott.com/media/5376/220614-cmnscp-initiation.pdf [Accessed January 31, 2023].
SRK Consulting (UK) Limited. 2020. Kgwakgwe Hill Manganese Project Independent Technical Report. https://minedocs. com/20/K-Hill-PEA-Feb-2020.pdf [Accessed January 19,2023].
Steenkamp, J.D., Chetty, D., Singh, A., Hockaday, S.A.C., Denton, G.M. 2020. From Ore Body to High Temperature Processing of Complex Ores: Manganese-A South African Perspective. https://doi.org/10.1007/s11837-020-04318-x
Steenkamp, J.D., Bam, W.G., Ringdalen, E., Mushwana, M., Hockaday, S.A.C., Sithole, N.A. 2018. Working towards an
increase in manganese ferroalloy production in South Africa - a research agenda. The Journal of the Southern African Institute of Mining and Metallurgy, vol. 118, pp. 645–654.
Tetra Tech Canada Inc. 2019. Technical report and preliminary economic assessment for the Chvaletice manganese project. https://www.miningnewsfeed.com/reports/Chvaletice_ PEA_03152019.pdf [Accessed January 21, 2023].
Tetra Tech Canada Inc. 2022. Technical Report and Feasibility Study for the Chvaletice Manganese Project, Czech Republic. https:// minedocs.com/19/Chvaletice-FS-07-27-2022.pdf [Accessed June 26,2023].
Tholana, T., Musingwini, C., Njowa, G. 2013. An algorithm to construct industry cost curves used in analyzing cash cost performance of operations for selected minerals in South Africa. The Journal of the Southern African Institute of Mining and Metallurgy. vol. 113, pp. 473–484.
Tsikos, H., Beukes, N.J., Moore, J.M., Harris, C. 2003. Deposition, Diagenesis, and Secondary Enrichment of Metals in the Paleoproterozoic Hotazel Iron Formation, Kalahari Manganese Field, South Africa. Economic Geology, vol. 98, pp. 1449–1462.
Van Zyl, H.J., Bam, W.G., Steenkamp, J.D. 2016. Identifying Barriers Faced by Key Role Players in the South African Manganese Industry. SAIIE27 Proceedings
Vanpée, R., Mayeres, I., Peduzzi, E. 2022. Study on the implications of EU policies for the affordability of car use in the future. https://www.fiaregion1.com/wp-content/uploads/2022/11/FIA_ Study_CostofOwnership.pdf [Accessed March 3, 2023].
Winjobi, O., Kelly, J.C. 2021. Update of the Manganese pathway in GREET® 2021. Argonne National Laboratory. https://greet.es.anl.gov/files/mn_update_2021 [Accessed February 2, 2023]. u
Affiliation:
Department of Industrial Engineering, Stellenbosch University, South Africa
Correspondence to:
G.M. Rohde
Email: rohdegustav1@gmail.com
Dates:
Received: 1 Jan. 2024
Revised: 19 Oct. 2024
Accepted: 25 Feb. 2025
Published: May 2025
How to cite:
Rohde, G.M., Bester, C. 2025. The potential of 4IR technologies to mitigate risk in mine residue management. Journal of the Southern African Institute of Mining and Metallurgy, vol. 125, no. 5, pp. 243–248
DOI ID: https://doi.org/10.17159/2411-9717/3536/2025
ORCiD:
G.M. Rohde
http://orcid.org/0009-0005-6174-5369
C. Bester
http://orcid.org/0000-0001-8967-050X
The potential of 4IR technologies to mitigate risk in mine residue management
by G.M. Rohde, C. Bester
Abstract
A major challenge in the mining sector is the responsible management and efficient storage of mine residue. Over recent years numerous tailings dams have experienced unintended spills and failures, with severe consequences for the mining industry, the environment, local communities, and the regional economy.
The fourth industrial revolution (4IR) is impacting society due to increased interconnectivity, processing speed, and automated technologies. This paper explores whether 4IR technologies have the potential to mitigate risks associated with tailings dams.
Experts in the field of tailings dam operational management were surveyed to determine their view on the risks associated with tailings dams and the potential of 4IR technologies to mitigate these risks. The survey found that a majority of experts believe that certain 4IR technologies have the potential to reduce the risks associated with tailings dam failures.
Keywords risk mitigation, 4IR technologies, tailings dam failures, tailings disasters
Introduction
The mining process is essential to produce goods, infrastructure, services, and to improve quality of life. Mines boost the economy in rural areas and provide job opportunities for the local communities; it contributes to the tax revenue required to support local municipalities and governments. Mining also has an immense impact on the environment. Estimates by Jawadand and Randive (2021) indicate that the production of 1 tonne of useable coal produces 0.4 tonnes of waste, and 1 tonne of useable copper generates 110 tonnes of waste.
A major challenge in the mining sector is the responsible management and efficient storage of waste resulting from the mining process (Aznar-Sánchez et al., 2018). In mining, tailings are the materials left behind after separating a valuable commodity from the uneconomical fraction (gangue) of the ore. Wet tailings are a sludge-like waste, comprising a mixture of water, chemicals, and fine gangue particles. This study focuses on wet tailings, which are commonly deposited in tailings dams. In the history of mining numerous tailings dams have collapsed with catastrophic monetary and human loss consequences (Owen et al., 2020).
Tailings dams represent some of the largest earth-fill structures in the world (Clarkson et al., 2021). There are an estimated 3 500 active tailings dam structures globally. An average of 2–5 major tailings dam failures and 35 minor failures occur per year (Caldwell et al., 2011). According to the literature, tailings dams fail significantly more than other constructed dams, such as water retention dams (Rana et al., 2022).
The fourth industrial revolution (4IR) is transforming many aspects of society (Schwab, 2017). The significant progression in interconnectivity and smart automation offers potential to improve the monitoring and management of tailings dams (Lumbroso et al., 2019). This could potentially improve tailings dam safety.
Tailings dam failures and associated risks
Traditionally tailings dams have been constructed based on design principles of water retention dams. Prior to the implementation of modern standards, tailings dams were designed less conservatively due to cost considerations (Vermeulen, 2001). The current practice in industry is to treat all aspects of operational management as site-specific (Vick, 1990; Dladla, Ramsamy, 2022). The cumulative tailings dam failure rate is as high as ~4.4%, which is much higher than the predicted failure rate of ~1.2% for normal water retention dams (Rana et al., 2022). A research study by Bowker and Chambers (2015)
The potential of 4IR technologies to mitigate risk in mine residue management
indicated that the number of general tailings dam failures decreased from the 1990s, but unfortunately the severity and number of fatalities increased (Bowker, Chambers, 2015).
As stated earlier, all tailings dam facilities are unique. Hence, the risks associated with each facility are site-specific, which makes risks difficult to generalise. However, the main ultimate failure mechanisms that have caused most failures, according to research studies, are overtopping, seepage, seismicity, and foundation failure (Bowker, Chambers, 2015; WISE, 2022; Piciullo et al., 2022), while others argue that overtopping and erosion are deemed the main failure modes (Mwanza et al., 2024). Extreme climatic events, mismanagement, and insufficient early detection may also be the root cause of dam failures. These issues are discussed in more detail in Table 1.
The 4IR and associated technologies
According to Klaus Schwab, the fourth industrial revolution (4IR) impacts industries, practices, and economies. It challenges and changes the way people think, operate, and make decisions. The revolution is characterised by a new era of digital driven technologies in the biological, digital, and physical world (Schwab, 2017).
Hermann et al (2016) summarised the fourth industrial revolution as a system that satisfies six design dimensions; these include visualisation, decentralisation, interoperability, real-time capability, modularity, as well as service orientation (Hermann et al., 2016). The 4IR further depends on a digital network and connections and includes Artificial Intelligence (AI), big data, Internet of Things (IoT), and machine learning (ML).
The 4IR includes numerous concepts and technologies, not all applicable to tailings dam monitoring and operations. The technologies most likely to be deployed on tailings dams are those associated with visualisation and integration of data and systems. These include augmented- and virtual-reality, IoT, machine learning, drone- and satellite-photography, and other concepts such as Light Detection and Ranging (LiDAR), and digital twins.
Survey design
For this research, a survey was developed and distributed to various experts in the field of tailings dam design, operation, and management in Sub-Saharan Africa. The main goal of the survey was to obtain opinions from experts and practitioners on the potential for 4IR technologies to better manage tailings dams, and which 4IR technologies have the highest potential to prevent failure and, thus, reduce risk.
It included experts employed at consulting firms, mining houses, tailings dam operators, and research professionals at
Root causes of tailings dam failure
Climate change
As stated by Azam and Li (2010), failures due to extreme weather events increased by 15% from the pre-2000s to the post 2000s. It is attributed to climate change with heavier storm events of shorter duration where the design of the facility is unable to contain and decant the storm event (Azam, Li, 2010).
universities. The distribution targeted an audience with a range of years of experience, ranging from less than five years to more than twenty years. The age and employment distribution ensured a spectrum of views.
There were 26 responses to the survey with relatively diverse profiles. The split between consultants and mining houses was 62% to 38%, respectively, with a homogeneous distribution of years of experience within the tailings dam industry.
Results
Tailings dam failure databases (Bowker, Chambers, 2015; WISE, 2022; Piciullo et al., 2022) were analysed to determine the ratio of documented failures that occurred from in-service versus decommissioned tailings dams. Overall, 82% of failures were reported on in-service tailings dams. This finding emphasises the importance of managing the risk associated with operations, monitoring, and management of in-service operational tailings dams.
Inherent risk rating
Experts were asked to rate the likelihood and consequence of the inherent risk associated with tailings dam failures. Although tailings dam failures are site specific, this approach combines the sitespecific risks to achieve a cumulative generalised perceived risk of failures. The likelihood ratings ranged from 0 (never) to 5 (almost certain), and the consequence ratings ranged from 0 (insignificant) to 5 (severe). The likelihood and consequence ratings, as perceived by experts, were used to determine an inherent risk rating for each of the failure mechanisms. The risk rating per survey for each failure mechanism was calculated using Equation 1 (Unguras et al., 2020).
Risk rating = likelihood x consequence [1]
The relative inherent risks by failure type as rated by the experts in the survey are shown in Figure 1. It becomes evident that the perceived risks associated with tailings dams range from ‘medium’ to ‘very high’.
According to Shah et al., risks rated this high require mitigation strategies, continuous monitoring, and management attention (Shah et al., 2020).
Current monitoring techniques used on tailings dams
Experts were asked to report on how tailings dams are currently being monitored, manually or automatic, and to what extent. For manual monitoring techniques, they were asked to assign a rating (0 = never manually monitored, to 5 = always manually monitored). Similar feedback was requested for automated monitoring techniques.
Mismanagement
Mismanagement of tailings facilities resulted in 20% increased dam failures from pre-to post 2000’s. This is attributed to operational guidelines not being followed, or rapid increases in mine production to benefit from a fluctuation in commodity prices. As a result, the rate of rise cannot be accommodated (Piciullo et al., 2022).
Insufficient early detection
Failures with causes such as seepage, overtopping, foundation failure, mismanagement, slope instability and erosion, typically show signs of deterioration prior to failure. The facility first reduces in stability, showing signs of irregularities before the failure occurs (Piciullo et al., 2022).
Table 1
The potential of 4IR technologies to mitigate risk in mine residue management
A weighted average was calculated by multiplying the assigned rating, 0 to 5, by the number of responses per rating. The average weighted rating was calculated for each manual and automated monitoring technique. Figure 2 illustrates the weighted averages, comparing the percentage difference between manual and automated monitoring adoptions per typical monitoring technique. It becomes evident that for all aspects monitored, manual monitoring currently outweighs the automated techniques.
Comparing all typical monitoring concepts associated with tailings dams, a weighting average of 3.4 was calculated for manual monitoring, and 2.1 for automated monitoring. It seems that manual monitoring techniques are 37% more likely to be adopted in industry than automated monitoring techniques.
A sigmoidal function (commonly referred to as an S-curve) is often used to describe the adoption of technology by industry. During the first phase, referred to as the initial or ‘emerging/ birth’ phase, adoption is slow. This is followed by an exponential growth phase where adoption occurs rapidly. The rate of adoption increases as benefits becomes clear. Finally, adoption decelerates during a phase referred to as the ‘mature and decline’ phase. This phase represents market and technology saturation (Denning, 2020; Cristóbal, 2017; Kucharavy, 2011).
The weighted average ratings that were used in Figure 2, were subsequently used in conjunction with the S-curve adoption rate theory. A zero-rating referring to no -, or minimum adoption in the market, and a five-rating refers to full adoption in the market and market saturation.
In Figure 3(a) and 3(b), the adoption level of both manual and automatic monitoring techniques was superimposed on adoption S-curves. It is evident that the majority of manual monitoring techniques are currently within the ‘mature’ phase. This is in contrast with the adoption level of autonomous monitoring techniques currently in the ‘growth’ phase. It is highly likely that the adoption of these techniques will grow and move to the exponential growth phase as benefits become clear and awareness spreads.
Figure 1—Perceived inherent risk matrix
Figure 2—Frequency of manual and automated techniques in industry
Figure 3—Manual and automated ratings overlayed on S-curves
The potential of 4IR technologies to mitigate risk in mine residue management
The potential of 4IR technologies to improve tailings dam operations
The top technologies that experts believe will improve tailings dam operations are: (1) IoT, (2) drone technology, and (3) machine learning. The potential of VR and AR are perceived to be significantly less.
A constraint of the survey is that the different technologies were rated in isolation. The integration of various technologies offers potential. This may be the reason IoT and machine learning, which are typically dependant on other information sources, were rated so high.
It is also evident from the rankings in Figure 4, that experts believe technologies associated with the integration of data and information (such as IoT and machine learning), can have a higher potential impact on understanding and anticipating tailings dams performance than other technologies used in isolation.
The potential application of the two highest rated technologies, IoT, and drone technology, as shown in Figure 4, is discussed in more detail in Table 2.
The potential of 4IR technologies to mitigate risks associated with tailings dams
The 4IR technologies shortlisted in this study offer varying potential to mitigate the different dam failure mechanisms. The survey respondents were asked to rate the potential of each 4IR technology to mitigate the risk of a specific failure mode. The judgement was captured on a 3-point scale (0 – no risk mitigation potential, 1 –some risk mitigation potential, 2 – high risk mitigation potential). The ratings (0, 1, or 2) of all the respondents were added together to get a holistic view of 4IR technologies risk mitigation potential. The number of respondents who believe that a specific technology has potential to improve tailings dam monitoring and operational management and can be used as a risk mitigation tool have been converted into a percentage and is shown in Figure 5.
The data depicted in Figure 5 imply that all failure mechanisms can potentially be anticipated by 4IR technologies. The experts indicated that machine learning and IoT are perceived to have the highest potential to mitigate risks associated with the failure mechanisms.
From the results it is also evident that different technologies seem to be better suited for different types of failure mechanisms. For example, drone technology seems to offer a high potential to mitigate risks associated with overtopping, and a low potential to mitigate risks associated with seismic events.
The four failure mechanisms that are best anticipated by 4IR technologies are: (1) overtopping, (2) slope instability, (3) seepage,
Potential of 4IR technologies to improve tailings dam performance
Internet of Things (Live Monitoring)
Drone Technology
Predictive Analytics
(Machine Learning)
Digital Twins
Satellite Photography
Lidar
Geospatial Technologies
High-Resolution
Atmospheric Forecasting
Augmented Reality (AR)
Virtual Reality (VR)
Figure 4—The relative potential of 4IR technologies to improve tailings dam performance
5—Potential for risk mitigation with 4IR technologies
and (4) external erosion. These occurrences are measurable or visible and therefore benefit from automated measuring techniques supported by the IoT or through drone surveys. The lowest mitigation potential of a failure mechanism by a 4IR technology are (1) structural inadequacies and (2) foundation inadequacies. These failure mechanisms could be more difficult to detect due to construction deficiencies and that their actual behaviour is hidden from physical observation. For example, a foundation failure due to weak spots in the foundation of the embankment, or a structural irregularity where the concrete in the penstock tower may crush due to poor concrete design or poor quality, is difficult to detect.
The potential conceptualised application of the two highest rated mitigation technologies
Drone Technology
In tailings dam management IoT can be utilised to connect and in realtime monitor the performance of a facility. The attributes measured and displayed in a central control room include all measurable aspects of the facility, such as seepage, pond sizes and volumes, and deposition- or solution volumes. These measurable items could also possibly come from other 4IR technologies such as drones, satellites, and sensors, linked to geospatial technologies.
Micro air unmanned vehicles have positively transformed the tailings dam industry in recent years. Experts believe that drones will have a significant impact on improving tailings dam operations in the future. Drones can be used on tailings dams to monitor operations on a continuous basis. These could include monitoring seepage, deformation, freeboard, pond area and volume measurements, beach slopes, and defects in areas previously inaccessible on foot.
Table 2
Internet of Things (IoT)
Figure
The potential of 4IR technologies to mitigate risk in mine residue management
Discussion and limitations
The survey results did not generate any surprising feedback and were mostly aligned with the literature review. Experts were overwhelmingly of the opinion that 4IR technologies could be used for risk mitigation of tailings dam operations. It might be questioned how reliably experts can judge the future potential of technology partially developed and currently in early stages of deployment. However, the survey results are consistent, and a degree of theoretical saturation was observed in the study. Saturation associated with quantitative research is reached when an increase in respondents does not seem to change the outcome of the survey.
Shortcomings were observed. Firstly, treating the failure mechanisms as independent occurrences is not necessarily correct, since some of the failure mechanisms are interlinked and difficult to analyse in isolation. The root causes and indirect triggers might be underappreciated and skewed as the survey only focused on the ultimate failure mechanisms. A second shortcoming of the survey is the fact that 4IR technologies were observed in isolation. 4IR technologies are often deployed in a system benefiting from the interconnection of several technologies.
An element that was not considered in this study was the cost and benefit of various technical options. The costs associated with 4IR technologies differ significantly (Chikwanda, 2023). An improvement and potential future extension to the study could be to quantify costs vs risk mitigation benefits per 4IR technology, and thus, consider their implementation.
Cybersecurity risks, a negative development associated with new technologies, have not been considered in the study. Information and data associated with tailings dams tend to be confidential and data leaks could possibly have an impact on the risks generated with the deployment of 4IR technologies. This might even result in sabotage or cyber ransom attacks.
The authors are of the opinion that the 4IR could have a substantial positive impact on operational monitoring and management of tailings dams. The interconnectivity, advanced analytics, and automated nature of these technologies should be very beneficial to mitigate risks associated with tailings dam failures.
Conclusions and recommendations
Research findings
Based on the experts’ ratings, the inherent risk of dam failure plotted in the medium to high-risk zone on a risk matrix, indicates that a significant number of risks are present in tailings dam operations.
Experts indicated that manual monitoring techniques are currently extensively used in monitoring dams. The manual monitoring techniques seem to be in the ‘mature’ phase of technology adoption. The newer automated techniques more recently deployed, are in the ‘birth and growth’ phases of adoption. It seems that a transition in the use of monitoring techniques might be underway from manual to automated techniques.
Although the survey was only completed by 26 respondents, all expressed a strong opinion that 4IR technologies can mitigate tailings dam risks. The development and propagation of all failure mechanisms can be mitigated by 4IR technologies, some offering higher potential than others. Overtopping and slope instability are the top two failure mechanisms that contribute to 48% of recorded failures. These two failure mechanisms have also been identified as benefiting most from the introduction of 4IR technologies. A 100% of respondents believe that drone technology could reduce the risk
of slope stability, whilst 96% of respondents believe that the IoT can reduce the risk of overtopping.
Table 3 summarises most of the findings of this research and the view of the surveyed experts. The type of dam failures is listed in the order of their perceived inherent risk. For each failure mechanism the performance criteria to monitor are listed and the most relevant 4IR technology and its likelihood of having the potential to mitigate risk is listed.
It is evident that 4IR technologies already contribute to the operation and management processes of tailings dams. The authors are of the opinion, and the survey also indicates, that 4IR technologies will significantly increase in use, efficiency, and accuracy in the years to come. The initial cost of 4IR technologies is limiting mining houses in implementing these technologies for tailings dams. It is however very likely that the costs of these technologies will reduce in the future. No single 4IR technology will solely be able to mitigate all risks associated with tailings dam failures, and a suite of integrated technologies should be considered per tailings dam to increase the efficiency of early, accurate, and efficient risk detection measures.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
References
Azam, S., Li, Q. 2010. Tailings Dam Failures: A Review of the Last One Hundered Years. Geotechnical News, vol. 28, no. 4, pp. 50–53. https://ksmproject.com/wp-content/ uploads/2017/08/Tailings-Dam-Failures-Last-100-yearsAzam2010.pdf
Aznar-Sánchez, J., García-Gómez, J., Velasco-Muñoz, J., Carretero-Gómez, A. 2018. Mining Waste and Its Sustainable Management: Advances in Worldwide Research. Minerals, vol. 8, no. 7, p. 284. https://doi.org/10.3390/min8070284
Bowker, L.N., Chambers, D. 2015. The Risk Public Liability & Economics of Tailings Facility Failures. US Department of Agriculture, vol. 15, no. 3, pp. 145–151. https://www.researchgate.net/publication/283321865
Caldwell, J.A., Van Zyl, D. 2011. Thirty Years of Tailings History from Tailings & Mine Waste. Tailings and Mine Waste 2010 https://doi-org/ezpcul/10.1201/b10569
Chikwanda, H. 2023. Fourth Industrial Revolution in the Vanadium Mining and Mineral Extraction Operations in South Africa. Sabinet African Journals, 2023, p. 12. https://hdl.handle. net/10520/ejc-aa_affrika1_v2023_nsi1_a11
Clarkson, L., Williams, D. 2021. An Overview of Conventional Tailings Dam Geotechnical Failure Mechanisms. Mining, Metallurgy & Exploration, vol. 38, pp. 1305–1328. https://doi.org/10.1007/s42461-021-00381-3
Cristóbal, J. 2017. The S-curve envelope as a tool for monitoring and control ofprojects. Procedia Computer Science, vol. 121, pp. 8–10. http://dx.doi.org/10.1016/j.procs.2017.11.097
Denning, P. 2020. The profession of IT: Technology adoption. Journal of the ACM, vol. 63, pp. 27–29. https://cacm.acm.org/ opinion/technology-adoption
Dladla, S.D., Ramsamy, S. 2022. Practical steps to Global Industry Standard on Tailings Management (GISTM) compliance for operational tailings storage facilities in South Africa. Journal of the Southern African Institute of Mining and Metallurgy, vol. 122, no. 6, p. 8. https://www.scielo.org.za/pdf/jsaimm/v122n6/06.pdf p. 8.
The potential of 4IR technologies to mitigate risk in mine residue management
Table 3
Risk mitigation model and proposed 4IR technologies
Failure Inherent Aspect to be Possible Mitigation Proposed mechanism risk monitored/mitigated solution potential technology
Slope stability, Continuos slope
Slope instability Very high
Overtopping Very high
Foundation High
Very high 96% 1st: loT (Live monitoring) seepage, general defects stability, early detection
Very high 88% 2nd: Satellite photography
Pond size, freeboard, Water management, early Very high 100% 1st: Drone technology weather forecasting detection, predictive analytics Very high 96% 2nd: loT (Live monitoring)
General foundation Continuous monitoring High 71% 1st: loT (Liver monitoring) inadequacies defects of foundation defects
Moderate 63% 2nd: Predictive analytics (Machine learning)/Digital twins
Seepage High Moisture on embankment
Moisture detection, Very high 92% 1st: loT (Live monitoring) perimeter continuous monitoring High 75% 2nd: Drone technology
General structural Continuous monitoring
Moderate 67% 1st: Predictive analytics inadequacies defects of structural defects (Machine learning)
Structural High
Seismic events Medium
External erosion Medium
Moderate 67% 2nd: loT (Live monitoring/ Digital twins)
Ground movement Ground movement detection, High 79% 1st: Predictive analytics acceleration warning signs, (Machine learning) predictive analytics
Moderate 63% 2nd: loT (Live monitoring)
Crack identification, Early detection, Very high 88% 1st: Satellite photography, crack dimensions continuous monitoring High 79% 2nd: loT (Live monitoring)
Denning, P. 2020. The profession of IT: Technology adoption. Journal of the ACM, vol. 63, pp. 27–29. https://cacm.acm.org/ opinion/technology-adoption
Hermann, M., Pentek, T., Otto, B. 2016. Design Principles for Industry 4.0 Scenarios. IEEE, pp. 3928–3937. https://ieeexplore. ieee.org/abstract/document/7427673
Jawadand, A., Randive, K. 2021. A Sustainable Approach to Transforming Mining Waste into Value-Added Products. Innovations and Sustainable Mining, pp. 1–20. https://doi.org/10.1007/978-3-030-73796-2_1
Kucharavy, D. 2011. Application of S-shaped curves. Procedia Engineering, vol. 9, pp. 559–572. https://doi.org/10.1016/j. proeng.2011.03.142
Lumbroso, D., McElroy, C., Goff, C., Collell, M.R., Petkovsek, G., & Wetton, M., 2019. The potential to reduce the risks posed by tailings dams using satellite-based information. International Journal of Disaster Risk Reduction, vol. 38, pp. 1–24. https://doi.org/10.1016/j.ijdrr.2019.101209
Mwanza, J., Mashumba, P., Telukdarie, A. 2024. A Framework for Monitoring Stability of Tailings Dams in Realtime Using Digital Twin Simulation and Machine Learning. Procedia Computer Science, vol. 232, pp. 2279–2288. https://doi.org/10.1016/j. procs.2024.02.047
Owen, J., Kemp, D., Lebre, E., Svobodova, K., & Perez Murillo, G. 2020. Catastrophic tailings dam failures and disaster risk disclosure. International Journal of Disaster Risk Reduction, vol. 42, pp. 1–10. https://doi.org/10.1016/j.ijdrr.2019.101361
Piciullo, L., Storrøsten, E.B., Liu, Z., Nadim, F., Lacasse, S. 2022. A new look at the statistics of tailings dam failures. Engineering Geology, vol. 303, p. 14. https://doi.org/10.1016/j. enggeo.2022.106657
Rana, N., Ghahramani, N., Evans, S., Small, A., Skermer, N., McDougall, A., Take, W. 2022. Global magnitude-frequency statistics of the failures and impacts of large water-retention dams and mine tailings impoundments. Earth-Science Reviews, vol. 232, pp. 1–26. https://doi.org/10.1016/j. earscirev.2022.104144, Schwab, K. 2017. The Fourth Industrial Revolution. Crown Currency. https://doi.org/KS/24869531
Shah, M.A.R., Renaud, F.G., Anderson, C.C., Wild, A. 2020. A review of hydro-meteorological hazard, vulnerability, and risk assessment frameworks and indicators in the context of naturebased solutions. International Journal of Disaster Risk Reduction, vol. 50, p. 12. https://doi.org/10.1016/j.ijdrr.2020.101728
Unguras, C.L., Anghelache, D., Vasilescu, V.G., Stoian, F., Ilcea, G. 2020. The generalized risk scale – a scalar integrated tool for developing risk criteria by consensus, in the field of explosives for civil uses. MATEC Web of Conferences, vol. 305, p. 78. http://dx.doi.org/10.1051/matecconf/202030500078
Vallero, D.A., Blight, G. 2019. Mine Waste: A Brief Overview of Origins, Quantities, and Methods of Storage. Waste (Second Edition), pp. 129–151. https://doi.org/10.1016/B978-0-12815060-3.00006-2
Vermeulen, N. 2001. The composition and state of gold tailings. Issue University of Pretoria, EBIT, pp. 2–7. https://doi.org/ CGT/369475821
Vick, S. 1990. Planning, Design, and Analysis of Tailings dams , BITech, Vancouver, B.C. Canada. https://doi.org/ EFAID/295381645
Williams, D. 2021. Lessons from Tailings Dam Failures—Where to Go from Here?. Minerals, vol. 11, no. 8, p. 853. https://doi.org/10.3390/min11080853
WISE, 2022. Wise Information Service on Energy. Uranium. https:// doi.org/WISE/395681254 u
Affiliation:
1Council for Scientific and Industrial Research, South Africa
2AYUDA Consulting Services, South Africa
3Anglo American Kumba Iron Ore, South Africa
Correspondence to:
T.J. Otto
Email: Theunis.otto@my-mine.co.za
Dates:
Received: 20 Aug. 2024
Revised: 19 Oct. 2024
Accepted: 24 Feb. 2025
Published: May 2025
How to cite:
Otto, T.J., Mkhatshwa, T., van Heerden, T.J., Cloete, C.H. 2025. Improving spatial mine-toplan compliance at an open pit mine through enhanced short-term mine planning. Journal of the Southern African Institute of Mining and Metallurgy, vol. 125, no. 5, pp. 249–258
DOI ID:
https://doi.org/10.17159/2411-9717/3546/2025
ORCiD:
T.J. Otto
http://orcid.org/0000-0002-3573-5627
Improving spatial mine-to-plan compliance at an open pit mine through enhanced short-term mine planning
by T.J. Otto1, T. Mkhatshwa2, T.J. van Heerden3, C.H. Cloete3
Abstract
The value realised by an open pit mine depends on the quality and integrity of the mine planning process as well as the level of execution against these mine plans. When managing the execution against the mine plan, how well the mine plan is executed spatially, is of critical importance. The short-term mine plan is the plan that is physically executed on most open pit mines. At the same time, the spatial mine-to-plan compliance is typically reconciled against the annual business plan. Short-term mine planning, therefore, plays an important role in the effectiveness of the spatial mine-to-plan compliance reconciliation process by providing operational teams with detailed designs and schedules, while ensuring that mining execution is spatially aligned with the business plan.
The Sishen open pit iron ore mine (Sishen) strives to continuously improve spatial compliance to the business plan. Sishen enhanced the role that short-term mine planning plays in enabling the forward-looking component of the spatial mine-to-plan compliance reconciliation process. The enhanced short-term mine planning process focuses on detailed tactical sequence designs per mining pushback, the health of value chain buffers, spatial plan-to-plan reconciliation, and the associated management routines.
This led to improved spatial mine-to-plan compliance to the business plan from 71% to 94% over the four-year period from 2020 to 2023 inclusive. These results indicate that the application of short-term mine planning as part of an integrated spatial mine-to-plan compliance process at open pit mines can contribute positively to improving the level of spatial execution against the business plan. This paper presents the enhancements made to the short-term mine planning process, which other open pit mining operations can consider to improve spatial mine-to-plan compliance.
Keywords
short-term mine planning, business plan, spatial mine-to-plan compliance, plan-to-plan reconciliation, detailed tactical sequence, management routines
Introduction
The economic objective of mining companies is to maximise the net present value (NPV) throughout the mine life in a sustainable way. The success of open pit mines is often determined by the mine’s ability to deliver actual value in line with the expected value that was ‘promised’ to investors and other stakeholders, based on mine plans. To ensure the sustainable success of a large open pit mine, two major areas need to be effectively managed. These are firstly, the quality and integrity of the mine planning process and secondly, the execution of the mine plan (Otto, 2019). The expected NPV of a mining operation is calculated using a mine plan as the technical basis. The actual value realised by a mining operation, is not only dependent on the quality and integrity of the mine plans. It also depends greatly on the level of execution against the mine plan. The level of execution or compliance against a mine plan can be measured in two ways: time-based (temporal) measurements, and area-based (spatial) measurements (Otto, Musingwini, 2020).
Spatial mine-to-plan compliance
Spatial mine-to-plan compliance can be defined as a measure tracking how well a mine plan is executed spatially. The level of spatial compliance to the mine plan is a key performance indicator (KPI) that is of critical importance to the open pit mining industry when managing the execution against the mine plan. Typical temporal mining reconciliation processes comparing tonnes, grade, and product alone do not adequately address the spatial aspect of open pit mining (Otto, 2019). According to Morley and
Improving spatial mine-to-plan compliance at an open pit mine
Arvidson (2017), the ultimate question to answer is: ‘Did we mine where and when we planned to mine?’ The effective measurement and management of spatial mine-to-plan compliance reconciliation is critical to answer this question. According to Otto (2019), this requires not only considering what is being mined when, but also where the mining activities are taking place in the open pit. It is important to identify where material has been mined to ensure that the progress of mine development is adequate to meet long-term strategic targets such as timely access to future ore targets (Hall, Hall, 2015). According to Esterhuysen (2013), the measurement and management of spatial mine-to-plan compliance is important to ensure:
➤ long-term sustainable supply of ore;
➤ that key issues negatively impacting on mine production are identified;
➤ the achievement of planned mining flexibility.
Otto and Musingwini (2019) provide insights into the major components and aspects associated with measuring and managing spatial mine-to-plan compliance at an open pit mine. Firstly, it is important to define ‘The Plan’ against which spatial mine-to-plan compliance is measured. Annual tactical plans (also called Business plans) are proposed as the basis for tracking the spatial mine-toplan compliance. Secondly, the spatial aspects of the actual areas mined should be determined. By comparing the latest surveyed digital terrain map (DTM) with the DTM developed at the start of the measuring period (month), areas where mining took place during the period under evaluation, can be identified. Thirdly, the spatial data is analysed to reconcile the actual mined areas with the planned areas in 3D space (Otto, Musingwini, 2019).
The actual areas mined, are spatially compared and evaluated to the areas planned in the annual tactical plan and then divided into four categories for reconciliation purposes. These categories are areas planned and not mined, areas mined out of sequence, areas mined out of plan and areas mined in plan. These categories are illustrated in Figure 1. The ‘planned not mined areas’ are those areas that were planned and not mined (indicated using the brown colour). The ‘mined out-of-sequence areas’ are those areas that were planned to be mined in the annual tactical plan, but they were mined out of their planned period. These areas are indicated by the yellow colour. The ‘mined out of plan areas’ are those areas that were not planned within the annual tactical plan but were mined (indicated by the red colour). The ‘mined in plan areas’ are those areas that were planned and mined. These areas are indicated in green.
Short-term mine planning
Mine planning is a mining engineering process that transforms a mineral resource into the best productive mining business (Morales, Rubio, 2010). The mine planning process is cyclical, and mine plans are regularly updated to incorporate the latest available information and changes in the macro-economic environment (Vivas, Nava, 2014).
According to Steffen (1997), most open pit mining operations follow a systematic and disciplined mine planning process involving three distinct levels of mine planning in developing the mineral reserves. These levels are the long-term (strategic), medium-term (tactical), and short-term (operational) mine planning horizons. Each of these horizons of mine planning represent different levels of risk and have different objectives. The planning horizons are nested in each other and the mining plans with a longer timeframe pass down guidance and restrictions in decisions to the shorter-term plans (Otto, Lindeque, 2021).
A mine plan, typically, has a design and a scheduling component. The design component provides the design for the mining activities such as the pit design, waste dump designs, ramp access designs, infrastructure requirements, and the design of mining activities within a mining bench. Developing an optimal mine design is a critical component of each of the mine planning horizons (Otto, Lindeque, 2021). The scheduling component considers the timing of these mining activities. For example, the design component will indicate the location and dimensions of an access ramp, and the scheduling component will indicate when the access ramp should be constructed.
Blom et al. (2019) discussed the major differences between long-term and short-term mine planning at open pit mines. Firstly, short-term plans are more practical and models mining activities at a higher level of detail when compared to long-term plans. Secondly, short-term plans cover a period of up to three months in daily and weekly increments, while long-term plans cover the life-of-mine using quarterly to yearly increments. Thirdly, short-term plans contribute to operational decision making while long-term plans provide inputs to strategic decision-making.
Short-term mine planning (STMP) is important because it links the guidance received from longer-term mining plans, and the execution thereof (Otto, Lindeque, 2021). Upadhyay and AskariNasab (2017) stated that the whole planning process at open pit mines is ineffective when STMP is not done well. An STMP is a rolling schedule that is updated weekly, fortnightly, or monthly, depending on the complexity and scale of the open pit mining operation. Short-term mine planning is typically conducted in detail on a day-to-day basis. The mining activities being considered in an STMP include access ramp construction, infrastructure establishment, block preparation, drilling, blasting, loading, and hauling (Otto, Lindeque, 2021).
The four main objectives for STMP are as follows (Blom et al., 2019; Blom et al., 2017; Burt et al., 2015; Upadhyay, Askari-Nasab, 2017; Vivas, Nava, 2014):
➤ Achieving the production throughput, and quality targets set by the tactical plans;
➤ The spatial alignment of the STMP to tactical plans;
➤ Enabling heavy mining equipment (HME) productivity;
➤ Practical executability.
When an open pit mine has an effective STMP process in place, accurate and detailed mine plans can be produced, communicated, and executed. In summary, STMP aims to ensure a detailed
Figure 1—Spatial mine-to-plan compliance categories (Otto, 2019)
Improving spatial mine-to-plan compliance at an open pit mine
understanding of the critical mining path, that will result in achieving the four objectives of throughput and quality, alignment with tactical plans, HME productivity, and executability (Otto, Lindeque, 2021). When developing an STMP it is critical that a balance should be found between spatial adherence to the tactical plan and the practical short-term realities in the open pit mine. The STMP links the tactical plans and the mining execution activities. The STMP (an output of the operational mine planning horizon) represents the ‘sharp end’ of the mine planning process as it is this plan that is physically executed (Otto, 2019). This paper explores how a well-developed STMP process supports the implementation of a spatial mine-to-plan compliance framework at an open pit mine. Examples from Sishen mine are included to illustrate the methodology.
Spatial mine-to-plan compliance framework
A high level of spatial compliance to the tactical mine plan is one of the major enablers for open pit mines to consistently, predictably, and sustainably deliver against budget targets and stakeholder expectations. The implementation of a comprehensive and integrated spatial mine-to-plan compliance reconciliation framework contributes to achieving high levels of spatial mine-toplan compliance. According to Otto and Musingwini (2019), the implementation of such a framework at an open pit mine firstly ensures that short-term KPIs are achieved by identifying key factors negatively impacting on mine production, and secondly, considers long-term KPIs such as sustainable ore supply and the achievement of planned mining flexibility.
Otto (2019) described the development and implementation of a spatial mine-to-plan compliance reconciliation framework at an open pit mine in South Africa. The framework enables mine planning engineers to proactively redirect mining (if required) using the operational mine planning horizon and to improve the executability of future mine plans. Figure 2 illustrates this spatial mine-to-plan compliance framework. It shows the six major components of the framework and the relationships between the components. The different components of the framework serve the following purposes (Otto, 2019):
➤ Components 1 to 3 provide guidance on defining and collecting data required to measure spatial mine-to-plan compliance.
➤ Component 4 utilises the value driver tree concept for drilling down of information and root cause analysis.
➤ Component 5 analyses the impact and consequence of the historic spatial mine-to-plan compliance on future compliance.
➤ The purpose of Component 6 is to determine the next best actions (NBAs) that are required to improve the spatial mineto-plan compliance.
➤ Feedback loops ensure the implementation of agreed corrective actions (also referred to as NBA) to improve the spatial mine-to-plan compliance. The first feedback loop enhances near real time visibility of the spatial progress of open pit mining and the second feedback loop facilitates improved tactical plans in future.
The green block identifies Component 6 where effective STMP (an output of the operational planning horizon referred to in the framework) plays a critical role. The NBA component is about applying the intelligence derived from the previous components of the framework for effective decision making.
The important role that STMP plays to define and implement the correct NBAs and ensure effective decision-making is highlighted by Otto (2019). Firstly, the STMP should actively seek to adhere to the spatial execution guidance provided by the tactical plan and direct mining to the correct spatial areas identified during the mine-to-plan compliance reconciliation process (Component 3 of the framework). Secondly, the STMP should incorporate identifying and actioning of priority tasks (PTs). PTs are focused actions that address the root causes of adverse spatial mine-toplan compliance performance identified in Component 4 of the framework. These PTs are defined as tasks that will enable the successful execution of the STMP. PTs are typically a list of key actions or key enablers linked to spatial areas of critical importance in the execution of the STMP and to the major inputs into the STMP. Ultimately, a reconciliation process adds value when it results in implementing corrective actions aimed at improving future performance (Morley, Arvidson, 2017).
Figure 2—Spatial mine-to-plan compliance framework (Otto, 2019)
Improving spatial mine-to-plan compliance at an open pit mine
Once the spatial alignment to the business plan (BP) is incorporated into the STMP, the open pit mine should seek to mine the blocks scheduled in the STMP and action the PTs. Executing the STMP then implies that the correct NBAs are being conducted. Effective decision making can now occur and will ensure that tactical planning outcomes are achieved in a sustainable manner. Effective STMP, therefore, underpins implementing a spatial mineto-plan compliance framework at open pit mines. STMP is the tool used to define and guide the implementation of the BP.
Sishen as a case study mine
The Sishen open pit iron ore mine (Sishen), is owned and operated by Anglo American Kumba Iron Ore (Kumba). It is located in the Northern Cape Province of the Republic of South Africa. Sishen is Kumba’s flagship mine and one of the largest open pit mines in the world. The mine consists of a single open pit with interconnected pushbacks stretching for approximately 14 km in the north-south direction (Kumba Iron Ore Limited, 2014).
Sishen extracts high-grade iron ore and the associated waste material utilising conventional truck and shovel open pit mining methods. The open pit mining method entails topsoil removal and stockpiling for later use during the waste dump rehabilitation process, followed by drilling and blasting of waste and ore. The waste material is in-pit dumped where such areas are available or hauled to waste rock dumps. The run-of-mine (ROM) iron ore is transported to the beneficiation plants, where it is crushed, screened, and beneficiated. The ROM iron ore is processed through two processing facilities namely a dense media separation (DMS) plant and a jig plant. The final product is transported via a railway line to the Saldanha harbour for subsequent export to various international markets.
The ex-pit material mined at Sishen is classified into iron ore and waste. In the 2023 financial year Sishen produced a total of 202.9 Mt ex-pit mined material comprising 39.1 Mt of ex-pit mined iron ore and 163.8 Mt of ex-pit mined waste material. The ratio of waste material to iron ore that needed to be extracted (stripping ratio) was at 4.2. In the 2023 financial year, Sishen produced a saleable product of 25.4 Mt (Kumba Iron Ore Limited, 2024). On
31 December 2023, Sishen declared an ore reserve of 599 Mt at an average quality of 54.1% Fe. The ore reserve was declared at a cutoff grade of 40% Fe and resulted in a reserve life of 15 years at the estimated production rate. The saleable product equated to 380 Mt at an average quality of 64.1% Fe (Moolman, Esterhuysen, 2023).
The 2023 Sishen reserve report has been prepared in accordance with the guidelines of the SAMREC Code.
Sishen typically operates in 10 active pushbacks, mining approximately 600 kt of ex-pit material per day. Figure 3 is a plan view showing Sishen’s active pushbacks targeted for ore and waste extraction and major waste dumps (WWD). Mining activities within the pushbacks are conducted using 12.5 m high mining benches. These activities are well coordinated to ensure safe execution and operational stability. The primary equipment deployed at Sishen includes primary blasthole drills (18 Pit viper 351D drills), rope shovels (3 P&H 4100 and 3 P&H 2800), hydraulic shovels (6 Liebherr R996), and trucks (36 Komatsu 960E haul trucks and 63 Komatsu 860E haul trucks). The P&H 4100 rope shovels are dedicated to loading pre-strip waste material to increase the capacity of waste material moved. The P&H 2800 rope shovels load internal waste and iron ore. The hydraulic shovels are assigned to loading buffer blocks, ramp establishment, and supporting the rope shovels (Mashala, 2023).
Sishen makes a substantial contribution to the operational and financial performance of Kumba, hence it is a significant operation to Kumba. The mine strives to consistently achieve budgeted production and financial targets while ensuring longterm sustainable iron ore supply. This objective is supported by implementing a spatial mine-to-plan compliance reconciliation process that ensures high levels of spatial compliance to the tactical mine plan. Effective STMP is a key enabler for successfully implementing the spatial mine-to-plan compliance reconciliation process in the open pit mining operations at Sishen. Therefore, STMP is key in improving the spatial mine-to-plan compliance reconciliation performance at Sishen.
Short-term mine planning at Sishen
The STMP process at Sishen is an established and structured mine planning process, which generates mine plans that are detailed, practically executable and, at the same time, aligned with the guidance provided from the BP, which is an output of the tactical planning horizon. Figure 4 is an overview of the mine planning process flow at Sishen. It illustrates how each planning horizon links to the next. It shows that the STMP (referred to as shortterm production strategy) plays a critical role in linking the BP or medium-term production strategy to work execution. The figure also indicates that the STMP comprises a mine design and a mine scheduling component (Gentle, 2021).
The STMP, at Sishen, covers a rolling period of three months or 12 weeks. The 12-week window provides clear visibility on the required mining production activities for that period, ensuring all the stakeholders are prepared in advance and aligned regarding priorities. The outputs of the STMP include the mining activities per bench, daily production deliverables such as ex-pit mining tonnages, ROM iron ore feed to the processing plants, and chemical qualities as well as PTs to be executed in support of the plan. The STMP is presented visually, which enables the different stakeholders to fully understand the interactions between mining activities captured in the plan. To continuously improve the quality and integrity of the mine plans and the alignment between the planning horizons, learnings from the STMP process are used as feedback or inputs into the next tactical planning cycle.
Figure 3—An overview of Sishen mine (Mashala, 2023)
Haul routes
Legend
Improving spatial mine-to-plan compliance at an open pit mine
One of the main objectives of the STMP at Sishen is to bridge the gap between the BP and the daily operational mining activities. Thus, ensuring that mining execution aligns with the BP, both from a temporal and a spatial perspective. The STMP provides the operational teams with detailed guidance on how the annual tactical plan should be executed. This includes specifying the blocks to be mined, allocation of resources, timelines allocated per activity, and the requirements for all the supporting priority tasks. The STMP, therefore, plays a critical role in directing the execution teams to the correct spatial mining areas and highlighting PTs that require operational focus, such as infrastructure placement, haulage ramp requirements, loading of sumps for water management, creation of drill accesses, and extension of power lines.
To support the role that STMP plays as a key enabler of spatial alignment between the BP and the daily mining activities, Sishen enhanced its STMP process during 2021 and 2022. The enhancements focused on improving four critical aspects of the STMP process namely STMP design, STMP scheduling, Plan-toPlan reconciliation, and the associated management routines.
In the STMP horizon at Sishen, the mine design component is referred to as the detailed tactical sequence (DTS). The DTS spans a period of three months and is updated quarterly. The main objective of a DTS is to design the mining sequence per pushback,
HME interaction, infrastructure considerations, access routes, and PTs prior to conducting the SMTP scheduling. The DTS aims to define the mining sequence and related short-term designs in such a way that it facilitates spatial compliance to the BP, optimises HME productivity, and ensures practical executability. Figure 5 shows an example of a DTS for Pushback 16 at Sishen. From the figure, the mining sequence design for Benches 14 and 13 can be seen. Firstly, the figure indicates the mining sequence and related HME interaction, including blocks to be drilled, blasted, and loaded. Secondly, it indicates infrastructure requirements such as electrical infrastructure, permanent ramp systems, and water infrastructure –critical enablers for the successful execution of the STMP. The visual nature of the DTS allows all relevant stakeholders to effectively understand and adhere to it.
The DTS provides highly visual 3D representations of how to effectively extract the required ex-pit material from a particular pushback by outlining the mining sequence and the interaction between different mining activities. It covers critical HME paths, including the access paths for drilling, loading, and hauling activities. It includes considerations for the interaction between the loading activities and other mining activities such as drilling, charging, blasting, highwall scaling, etc. Additionally, the DTS incorporates PTs that support the execution of the STMP such as infrastructure placement, access ramp requirements, and surface water management. Sishen introduced the concept of rain readiness per pushback as a PT, thereby formalising and prioritising surface water management as a critical component of the DTS designs. Figure 6 is a visual representation of the key aspects associated with rain readiness in the Pushback 16 area. The figure identifies the positioning of sumps and surface waste pipe infrastructure and facilitates discussion around the establishment of the rain readiness infrastructure and the impact on other mining activities in the pushback.
Developing DTS designs improves the practical executability of the STMP, leading to improved spatial compliance to the BP because risks and opportunities associated with the mining design per pushback have been identified prior to the STMP scheduling process. The DTS serves as an essential input into the scheduling portion of the STMP process. At Sishen, this is referred to as the 12-week schedule.
The 12-week schedule is a time-based component of the STMP. It transforms the DTS designs into a time-based sequence of interrelated and interdependent mining activities and ensures that
Figure 4—Overview of the mine planning process at Sishen (Gentle, 2021)
Figure 5—Detailed tactical sequence of Pushback 16 at Sishen (Mantsena, 2023)
Improving spatial mine-to-plan compliance at an open pit mine
the tonnage and ROM iron ore quality results from the different pushbacks effectively combine to produce the required ex-pit mining output at a mine level. In addition to the DTS designs, the 12-week schedule incorporates cross-functional inputs such as HME maintenance considerations, HME performance assumptions, and processing plant inputs.
To ensure stability within the mining operation at Sishen, the 12-week schedule actively targets and effectively manages the health of mining value chain buffers, such as the number of blocks staked, the number of blocks being drilled, and the amount of blasted material. ROM stockpiles (another buffer in the mining value chain) are also managed through the STMP to ensure that the feed to the processing plants is aligned with BP expectations. Having these buffers in place enables the mine to effectively deal with short-term variability without the need to deviate from the 12-week schedule, thus, improving spatial mine-to-plan compliance. Figure 7 illustrates how the drilled and blasted floor stock, high-grade run-of-mine (ROM) iron ore stockpile level, and finished product stockpile levels are tracked on a monthly basis to ensure that these buffers stay at levels above the planned levels required.
PTs, identified and agreed for execution as part of the DTS designs, are incorporated in the 12-week schedule. This process ensures that the PTs are assigned (with due dates and KPIs to measure successful completion) to the responsible team members for the mining areas where actions are required. The incorporation of PTs in the 12-week schedule leads to improved spatial mine-toplan compliance because these enablers to the spatial deployment of the open pit mine are actively managed.
The outputs of the 12-week schedule are a visual representation of mining activities. An example is shown in Figure 8. The figure shows the timing of mining activities such as block preparation, drilling, and loading. Utilising 3D visualisations to represent the 12week schedule output, enables stakeholders to visualise the mining sequence and associated HME interactions, allowing them to fully understand risks, such as maintaining adequate blasted stock, and contribute towards optimising the final agreed schedule (Gentle, 2021).
As indicated before, the STMP is the mining plan that is physically executed by the mining operations team. It is, therefore, critical that the spatial mining areas targeted by the STMP align with the mining areas targeted by the BP. At Sishen, this alignment is validated through the introduction of a spatial plan-to-plan reconciliation process, which compares how well the STMP aligns with the tactical plan spatially.
Having a detailed, well-designed, and highly executable STMP is only one half of the requirement. For the STMP to be effective and ensure that “the right things are done right”, the STMP has to also align spatially to the BP. At Sishen, the STMP actively seeks to align spatial deployment priorities with the BP and actively limits the mining of areas outside of the footprint defined by the BP. To validate that the STMP consistently aligns spatially to the BP per pushback, the period progress plots (or stage plans) of the STMP are compared to the stage plans of the BP monthly. This process is referred to as spatial plan-to-plan reconciliation. The objective of this spatial reconciliation is to ensure that the mining areas targeted by the STMP, and therefore executed by the mining operations team, are aligned with the spatial deployment guidance from the BP. The process also highlights the areas that are scheduled to be mined in the BP but have not been scheduled in the STMP. Should these yet unmined areas require additional focus, they are then assigned to the PT list. According to Otto and Musingwini (2020), measuring the spatial plan-to-plan reconciliation gives the operation a forward-looking view on the likelihood of achieving their spatial mine-to-plan targets against the BP.
Figure 9 is a view of the output from a plan-to-plan reconciliation conducted for Pushbacks 16 at Sishen. The figure visually compares the areas planned in the STMP with the areas planned in the BP. The colours used to represent the spatial reconciliation categories are aligned with the categories discussed earlier in the paper and illustrated by Figure 1. It can be seen from the figure that there are areas that are spatially aligning with the BP (areas indicated in green), there are areas that are planned to be mined in the annual tactical plan but not yet scheduled in the STMP (areas indicated in brown), and there are areas that are scheduled to be mined in the STMP, which are outside of the annual tactical plan (areas indicated in red).
Figure 6—Key components of rain readiness planning in Pushback 16 (Cloete, 2022)
Figure 7—Mine value chain buffers (Mashala, 2023)
Improving spatial mine-to-plan compliance at an open pit mine
The results from this reconciliation are used as inputs into the next STMP cycle. Mining is directed to the ‘brown areas’ and PTs required to enable mining in these areas, if any are identified, while the necessity of mining in the ‘red areas’ is debated and escalated for approval as required. This feedback loop continuously improves the quality of the STMP.
At the core of a successful STMP process lies integration, collaboration, and alignment (Otto, Lindeque, 2021). Sishen has improved the collaboration between the mining operations and mine technical services (MTS) teams by introducing formalised management routines supporting the development and approval of the monthly STMP. These routines are embedded through four formalised meetings, namely a monthly STMP inputs meeting, a monthly STMP review meeting, a monthly STMP sign-off meeting, and weekly pit visits meetings. These management routines ensure the planning process is predictable, repeatable and effective, and improve the quality of the STMP.
The STMP inputs meeting is held during the first week of every month and kicks off the monthly process that leads to the sign-off of the next STMP for execution. The purpose of this meeting is to agree on the terms of reference (or major input assumptions) that will form the basis of the next cycle of the STMP process. The meeting ensures that all key inputs are adequately analysed, discussed, and agreed by process owners prior to them being included into the STMP scheduling process. Importantly, the inputs are also evaluated against the BP to ensure alignment.
The agenda of the meeting typically covers the following: a review of the previous month’s spatial mine-to-plan reconciliation performance, key input assumptions regarding HME performance, haul route and waste dumping inputs, grade control, product quality and processing plant assumptions, geotechnical considerations, and rain readiness aspects such as the position of sumps and areas of anticipated water accumulation. The DTS designs and PTs are also reviewed per mining pushback.
The meeting is chaired by the MTS manager, whilst representatives from the mining operations teams are led by the Manager Mining. The involvement of senior mine leadership enforces the importance of the meeting, creates a sense of discipline in terms of attendance, and ensures that quality decisions are taken to provide clear direction to the team developing the STMP. All relevant stakeholders involved in mining operations and MTS disciplines attend this meeting.
All risks and opportunities associated with the STMP, as well as ways to mitigate those risks and capitalise on the opportunities, are discussed in this meeting. If a topic is not raised in the STMP inputs meeting it is unlikely to form part of the STMP. Ultimately, the meeting provides clear direction to the team developing the STMP and prevents re-work. The STMP team can now effectively and efficiently complete the STMP schedule, which is then reviewed in the second meeting, namely the STMP review meeting.
The STMP review meeting occurs in the second week of the month. The objective of this meeting is to review the temporal and spatial outputs from the STMP schedule with the aim of firstly, ensuring alignment with the BP expectations and secondly, ensuring that the STMP schedule is effectively enabling the execution of the DTS designs and PTs per mining pushback. The meeting requires critical analysis of the 12-week schedule by all key stakeholders and process owners to adequately assess, challenge, and optimise the schedule prior to execution. As a part of the discussion, the following key questions are evaluated (Gentle, 2021):
➤ Does the schedule meet temporal and spatial annual BP targets?
➤ Does the sequence follow the DTS?
➤ How are drilling and blasting activities and the associated interaction with HME addressed?
➤ Have haul route, waste dump, and ROM stockpile options been adequately considered and incorporated in the 12-week schedule?
Figure 8—12-Week schedule output of Pushback 16 at Sishen (Moeng, 2023)
Figure 9—STMP versus BP plan-to-plan compliance for Pushback 16 at Sishen (Pilane, 2023)
Improving spatial mine-to-plan compliance at an open pit mine
➤ What is the status of agreed mining value chain buffers?
➤ Does the 12-week schedule incorporate the agreed PTs, including assigning responsibilities and due dates for completion?
The collaborative nature of the meeting ensures that the STMP is rigorously evaluated and optimised. It provides a high degree of confidence in the execution of the STMP to all stakeholders, thus, ensuring practical executability of the STMP leading to improved spatial compliance to the BP. Following the STMP review meeting, the 12-week schedule is finalised for formal sign-off and communication.
The third management routine is the STMP sign-off meeting, which occurs in the third week of the month. In the signoff meeting the STMP is physically signed off by the relevant stakeholders as a formal way of officially approving the monthly STMP and committing to achieving the requirements.
To ensure continuous alignment between the MTS team and the mining operational team, the STMP is discussed every Monday in the pit during a weekly pit visit meeting. This is the fourth meeting to complete the STMP management routines. The focus of the discussion is progress against the planned mining activities (including block preparation, drilling, blasting, loading, and hauling) for the current week and the following week. Progress with PTs is also tracked and discussed. The value of conducting this meeting face to face and in the pit is that team members have the same context of unforeseen risks and opportunities experienced on a day-to-day basis by the mining operational team. Now the teams can directly compare the planned versus actual mining activities. Being in the pit, the team members can visually inspect the mining progress, allowing for real-time assessment of how closely actual mining activities align with the STMP. Deviations that are picked up can be immediately addressed, allowing the mining operations to maintain their compliance to the STMP plan. These meetings solidify alignment between the mining operations team and the MTS team. Learnings from these meetings are also fed into the inputs meeting of the next STMP cycle.
In summary, the management routines followed at Sishen create a platform for collaboration between the teams developing the STMP and the teams executing the STMP. These interactions improve the quality of the STMP as well as the level of spatial execution of the STMP. The enhancements of the other components of the STMP process at Sishen enable and validate the spatial alignment between the STMP and the BP. In combination, these positively impacted on Sishen’s spatial mine-to-plan compliance reconciliation performance.
Impact on spatial mine-to-plan compliance
At Sishen, the spatial alignment between mining execution and the BP is measured through a spatial mine-to-plan compliance reconciliation process. Otto and Musingwini (2019) discussed the implementation of spatial mine-to-plan compliance reconciliation at Sishen in the period between 2014 and 2016 and highlighted that the approach incorporated a model providing mine-to-plan reconciliation categories as well as a spatial reconciliation process focusing on reporting, target setting, and analysis of the reasons for deviations. At the time, the application of STMP as a tool to define and guide the spatial implementation of the BP (by bridging the gap between the BP and the daily operational mining activities and ensuring that spatial mining execution aligns with the BP) was not adequately incorporated in the spatial reconciliation process at Sishen. This contributed to a regression in the spatial mine-to-plan
compliance reconciliation results after 2016. This paper provides insights into how the incorporation of effective STMP, in the spatial mine-to-plan compliance reconciliation process at Sishen during 2019 and 2020, led to the mine regaining the targeted spatial mineto-plan compliance reconciliation results at the end of 2022 and sustaining these results into 2023.
Spatial mine-to-plan compliance against the BP is measured, reconciled, and reported monthly. The reporting format includes graphs in which the spatial mine-to-plan compliance is expressed as a percentage of planned tonnes mined using the categories suggested by Otto and Musingwini (2019). Compliance measurement per category is calculated as a tonnage compliance and expressed as a percentage of planned tonnes for ore, waste, and total material mined. The spatial mine-to-plan reconciliation results are compiled and expressed cumulatively over a 12-month period. Figure 10 is an example of a graph used to report the spatial mine-to-plan compliance at Sishen for the open pit mine as a whole. The figure illustrates that the cumulative spatial mine-to-plan compliance, for the year featured in the graph, was 82% at the end of May and increased to 94% by the end of December. This is relative to a target of 85% at Sishen.
Sishen also reports the spatial mine-to-plan compliance reconciliation results per pushback. This additional level of reporting provides visibility on the spatial reconciliation performance per individual mining area. Using these insights into the spatial reconciliation ‘health’ of the pushbacks assists in defining which pushbacks should receive more focused attention during the development of the next STMP, with the aim of improving the spatial mine-to-plan compliance performance of that pushback. Figure 11 is a graph indicating how the spatial mine-to-plan compliance is reported per pushback. From the graph it can be seen that Pushbacks 1, 8.1, 8.2, and 13 have significant (higher than 5%) planned areas that were not mined. Pushbacks 1 and 8.2 also present a high percentage of areas mined out-of-sequence. To address this, during the development of the next STMP, focus should be given to prioritise inputs, which will mitigate the challenges encountered in these pushbacks with the aim of improving the spatial compliance going forward. This could include redirecting mining capacity from mining areas ahead of plan towards mining the planned not mined areas. For the other pushbacks it is important to maintain the performance while ensuring adherence to the BP.
As discussed before, the results from the spatial mine-to-plan compliance reconciliation process are incorporated into the STMP routines. The reconciliation results are discussed and thoroughly analysed during the monthly STMP inputs meeting. The focus is firstly, on understanding the reasons for not mining the planned areas and for mining the areas outside of the annual BP (if any) and secondly, on identifying those pushbacks requiring specific interventions to address unsatisfactory spatial mine-to-plan compliance performance. These discussions ensure that the STMP directs mining activities to the correct spatial areas and incorporates the identified PTs.
The enhanced STMP process at Sishen has positively impacted the spatial mine-to-plan compliance performance against the BP. Figure 12 shows the end-of-year (December) spatial mineto-plan compliance graphs from 2020 to 2023 inclusive. From the information presented in Figure 12, the spatial compliance at Sishen improved from 2020 to 2023. At the end of 2020 the spatial mineto-plan compliance, measured against the approved BP at Sishen, was at 71%. At the end of 2022 the spatial mine-to-plan compliance had improved to 94% against a target of 85%. This performance was
Improving spatial mine-to-plan compliance at an open pit mine
repeated in 2023, illustrating the sustainability of the results. The figure also indicates a reduction in the areas planned and not mined from 25% in 2020 to 6% in 2023, and a reduction in the areas mined outside of the annual BP, from 4% in 2020 to 0% (no mining outside of the annual BP) in 2023.
Conclusion
This paper presented the role that STMP plays in ensuring the effectiveness of the spatial mine-to-plan compliance reconciliation process at an open pit mine. A well-developed STMP process improves the alignment between mining execution and the BP, both from a temporal and a spatial perspective. An effective STMP process bridges the gap between the BP and the daily operational mining activities. The STMP is the plan that is physically executed and thus, provides the operational teams with detailed guidance needed to ensure mining execution is aligned with the BP requirements.
STMP is the ideal tool to use to facilitate and enable the forward-looking component of the spatial mine-to-plan compliance reconciliation process at an open pit mine. At Sishen, it is used to define and implement the corrective actions resulting from this reconciliation process, thus ensuring effective decision making.
Enhancements were made to the STMP process at Sishen, aimed at supporting the spatial mine-to-plan compliance reconciliation process. Firstly, introducing the DTS designs improved the practical
executability of the STMP, which led to improved spatial compliance with the BP. The sequence per pushback, HME interaction, infrastructure considerations, access routes, and PTs are evaluated prior to the STMP scheduling process. Secondly, the STMP targets and effectively manages the health of mining value chain buffers. Focusing on these buffers enables the mine to effectively deal with short-term variability without the need to deviate from the STMP, thus, improving spatial mine-to-plan compliance. Thirdly, the introduction of a spatial plan-to-plan reconciliation, which compares the areas planned for mining in the STMP with those targeted in the BP confirms spatial alignment between the mine planning horizons. Finally, formal management routines, supported by senior mine leadership, are used to improve the collaboration between the mining operations and MTS teams. These management routines ensure that the STMP process is effective and improve the quality of the STMP.
These enhancements in the STMP process contributed to an improvement in the spatial mine-to-plan compliance reconciliation performance at Sishen in the period between 2020 and 2023 inclusive. The spatial areas mined within the approved annual BP increased from 71% at the end of 2020 to 94% at the end of 2022 and this performance was repeated in 2023, illustrating the sustainability of the results. In addition to the improved spatial mine-to-plan compliance reconciliation results, the increased alignment between the BP and mining execution contributed
Figure 10—Spatial mine-to-plan compliance, Sishen performance (Pilane, 2022)
Figure 11—Spatial mine-to-plan compliance per pushback, Sishen (Pilane, 2022)
Improving spatial mine-to-plan compliance at an open pit mine
12—Sishen end-of-year spatial mine-to-plan compliance reconciliation results
positively to the mine safely achieving production targets in line with the BP expectations. Effective STMP supported the implementation of a spatial mine-to-plan compliance reconciliation process at Sishen and is a key contributor to the mine delivering and improving spatial mine-to-plan compliance to the BP. The approach presented in this paper can be adapted by other open pit mining operations to improve alignment between mining execution and the BP.
Acknowledgements
The permission granted by the management of Anglo American Kumba Iron Ore to publish the paper is greatly acknowledged.
References
Blom, M., Pearce, A.R., Stuckey, P.J. 2017. Short-term scheduling of an open pit mine with multiple objectives. Engineering Optimization, vol. 49, no. 5, p. 777–795.
Blom, M., Pearce, A.R., Stuckey, P.J. 2019. Short-term planning for open pit mines: a review. International Journal of Mining, Reclamation and Environment, vol. 33, no. 5, pp. 318–339. DOI: 10.1080/17480930.2018.1448248
Burt, C.N., Lipovetzky, N., Pearce, A.R., Stuckey, P.J. 2015. Scheduling with Fixed Maintenance, Shared Resources and Nonlinear Feedrate Constraints: A Mine Planning Case Study, CPAIOR 2015, LNCS 9075, pp. 91–107, 2015. Cloete, C.H. 2022. Sishen 2022/23 Rain Readiness Plan Development. Internal Kumba presentation, August 2022, pp. 1–12.
Esterhuysen, W.D. 2013. Progress Update on Mine Planning Reconciliation in Kumba. Internal Kumba Exco memo, June 2013, pp. 1–8.
Gentle, C. 2021. Best Practice Principles: Detailed Tactical Sequence Version 1.0. Internal Anglo American standard, 18 March 2021, pp. 1–12.
Hall, A., Hall, B. 2015. The role of mine planning in high performance. Bulletin Magazine, August 2015, pp. 68–71. Kumba Iron Ore Limited. 2014. Sishen Mine SEAT report 2014. https://www.angloamericankumba.com/~/media/Files/A/ Anglo-American-Group-v5/Kumba/media/publications/ sisshen-seat-final-2feb.pdf , pp 1–85, [Accessed 5 July 2024].
Kumba Iron Ore Limited. 2024. Annual Financial Statements 2023. https://www.angloamericankumba.com/~/media/ Files/A/Anglo-American-Group-v5/Kumba/investors/annualreporting/annual-financial-statements-2023.pdf, 29 February 2024, pp. 1–140, [Accessed 9 July 2024].
Mantsena, R. 2023. STP Sign off document, Sishen. Internal Kumba presentation, June 2023, pp. 1–184.
Mashala, K. 2023. 2023-2027 Business plan presentation. Internal Kumba report, November 2023, pp 1–133.
Moeng, K. 2023. Quarterly and 18 Month MAD & Detail Tactical Sequence. Internal Kumba presentation, March 2023, pp. 1–66.
Moolman, I., Esterhuysen, W.D. 2023. Sishen mine: Ore reserve (and saleable product) statement 2023, 31 December 2023, pp. 1–134.
Morales, C., Rubio, E. 2010. Development of a mathematical programming model to support the planning of short-term mining. APCOM 2010, pp. 3–15.
Morley, C., Arvidson, H. 2017. Mine value chain reconciliation –demonstrating value through best practice. Tenth International Mining Geology Conference, Hobart, Tasmania, 20–22 September 2017, pp. 279–292.
Otto, T.J. 2019. A spatial mine-to-plan compliance framework for open pit iron ore mines, PhD Thesis, University of the Witwatersrand, Johannesburg, South Africa, pp. 1 – 162.
Otto, T.J., Lindeque, G.C. 2021. Improving productivity at an open pit mine through enhanced short-term mine planning. The Journal of The Southern African Institute of Mining and Metallurgy, vol. 121, no. 11, pp. 589–598.
Otto, T.J., Musingwini, C. 2019. A spatial mine-to-plan compliance approach to improve alignment of short-term and long-term mine planning at open pit mines. The Journal of The Southern African Institute of Mining and Metallurgy, vol. 119, no. 3, pp. 253–259.
Otto, T.J., Musingwini, C. 2020. A compliance driver tree (CDT) based approach for improving the alignment of spatial and intertemporal execution with mine planning at open pit mines. Resources Policy, vol. 69. https://doi.org/10.1016/j. resourpol.2020.101826
Pilane, T.2022. Mine-to-plan compliance presentation, Sishen Mine. Internal Kumba report, December 2022, pp. 1–60.
Pilane, T. 2023. Mine-to-plan compliance presentation, Sishen Mine. Internal Kumba report, December 2023, pp. 1–62.
Upadhyay, S.P., Askari-Nasab, H. 2017. Dynamic shovel allocation approach to short-term production planning in open pit mines. International Journal of Mining, Reclamation and Environment, pp. 1–20.
Steffen, O.K.H. 1997. Planning of open pit mines on a risk basis. Journal of the South African Institute of Mining and Metallurgy, vol. 97, March/April 1997. pp. 47–56.
Vivas, R.E., Nava, A. 2014. An Integrated Mine Plan - Connecting Long, Medium- and Short-Term Planning Strategies at Goldcorp Penasquito. Mining Engineering Magazine, vol. 66, no. 12, pp. 29–34. u
Figure
Affiliation:
Chemical Resource Beneficiation (CRB), Hydrometallurgy Group, North-West University, South Africa
Correspondence to:
D.J. van der Westhuizen
Email: derik.vanderwesthuizen@nwu.ac.za
Dates:
Received: 15 Oct. 2025
Published: May 2025
How to cite: Kruger, M., Krieg, H., van der Westhuizen, D. 2025. Separation and recovery of cobalt and aluminium from spent gas-to-liquid catalysts using CYANEX® 272. Journal of the Southern African Institute of Mining and Metallurgy, vol. 125, no. 5, pp. 259–266
DOI ID:
https://doi.org/10.17159/2411-9717/713/2025
ORCiD:
M. Kruger
http://orcid.org/0000-0002-1919-068X
H. Krieg
http://orcid.org/0000-0002-4268-133X
D. van der Westhuizen
http://orcid.org/0000-0001-6764-4132
This paper is based on a presentation given at the Hydrometallurgy Conference 2024 1-3 September 2024, Hazendal Wine Estate, Stellenbosch, Western Cape, South Africa
Separation and recovery of cobalt and aluminium from spent gas-to-liquid catalysts using CYANEX® 272
by M. Kruger, H. Krieg, D. van der Westhuizen
Abstract
In the field of Fischer-Tropsch synthesis catalysts, cobalt, platinum, and aluminium are the primary metal constituents of the conventionally used cobalt catalyst. While platinum recovery remains a focal point in this field, a recent endeavour by a South African company emerged, which aimed to recover cobalt and aluminium from spent Fischer-Tropsch synthesis catalysts for potential use in the agricultural sector. This study aimed to optimally separate cobalt and aluminium from spent Fischer-Tropsch synthesis catalyst for this purpose. It further leverages the OLI database to predict metal speciation during solvent extraction to aid with experimental planning. Shake-out tests were conducted to determine optimal separation conditions, which were validated using three distinct methods: (i) evaluating the experimental error, (ii) comparing results with similar research, and (iii) validating results against the OLI database. The investigation revealed the pivotal role of the aqueous pH yielding effective separation at pH 3.13 ([H2SO4] = 2.5 × 10−4 M) when using 20 vol.% CYANEX 272 and 50 mol.% pre-neutralisation. Effective cobalt scrubbing was achieved with 50 g/L Al at pH 2.8, while successful stripping required 1 M H2SO4. Competitive interactions between complexing aluminium and cobalt species were observed when contacting with CYANEX 272 at a pH < 4. Equipment design analysis for a targeted separation efficiency of 87% dictated the necessity of two mixer stages and a settler measuring 1.5 m × 1.5 m × 6.0 m (height × width × length). This settler would ensure sufficient residence time for gravity separation at a flow rate of 10 m3/h.
Keywords Fischer-Tropsch synthesis, solvent extraction, CYANEX 272, OLI, metal speciation
Introduction
In 1922, the Fischer-Tropsch synthesis (FTS), formulated by Franz Fischer and Hans Tropsch, emerged as a pivotal process for the conversion of synthetic gas (syngas), primarily consisting of carbon monoxide (CO) and hydrogen (H2), into diverse hydrocarbons such as oil, petrol, diesel, and various other chemical compounds. Integral to this transformative reaction is its catalyst. Traditionally, catalysts rooted in ruthenium, nickel, iron, and cobalt have served in the FTS process (Jahangiri et al., 2014). Cobaltbased catalysts stand out as the preferred option within this spectrum due to their cost-effectiveness, heightened activity, selectivity, and extended operational lifespan (Liang et al., 2019). Typically, these cobalt-based catalysts consist of approximately 10 mass% cobalt (Co), < 2 mass% platinum (Pt), with the remaining composition commonly made up of an Al2O3-based supporting material (Jacobs et al., 2002).
Like all catalysts, these components degrade over time. Upon reaching a deactivation threshold (inadequate activity for sustaining the desired productivity), the spent catalyst is replaced with a fresh batch. Presently, Minemet, a company situated in South Africa, specialises in reclaiming the Pt from the depleted catalyst. However, considering the potential agricultural advantages of Co and aluminium (Al) – Co enhances crop yields and mitigates vitamin B12 deficiencies in livestock, while Al alters soil pH, facilitating the growth of plants accustomed to acidic soil conditions (International Plant Nutrition Institute, 2014), it would be beneficial to also recover these metals. Consequently, in this study, the aim was to effectively recover Co and Al from spent FTS catalysts for their agricultural applications. Two final product specifications could be targeted: (i) Al-based (< 100 ppm of Co) or (ii) Co-based (3:1 massbased ratio of Co to Al), of which the Al-based product was targeted in this study.
Fortunately, many methods have been developed to separate these metals. Within the field of metallurgy, hydrometallurgy focuses on employing processes using aqueous solutions to extract
Separation and recovery of cobalt and aluminium from spent gas-to-liquid catalysts using CYANEX®
or separate metals. Various techniques have been used for the separation of Co and Al, including solvent extraction (SX) (Tsakiridis and Agatzini-Leonardou, 2005), ion exchange (IX) (Botelho et al., 2019), precipitation (Dhiman, Gupta, 2019), selective leaching (Chong et al., 2013), and crystallisation (Ferreira et al., 2009). However, none of these studies have focused on spent FTS catalysts, which contain different metal compositions, thus requiring the development of different optimal separation conditions.
Among these hydrometallurgical techniques, SX has emerged as the preferred method compared with leaching, IX, and crystallisation, due to its remarkable selectivity towards specific metals (Sole, 2008). Furthermore, SX was shown to be effective and feasible for the separation of Co and Al (Tsakiridis, AgatziniLeonardou, 2005). Hence, in this study, SX was selected as the technique to separate Co and Al. The efficiency of the separation process can be quantified by determining the extraction efficiency (E) and the distribution coefficient (D), as described in Equations 1 and 2, respectively:
where E refers to extraction/separation efficiency (%), m refers to the mass of metal i in the organic phase (org) after contact, and in the aqueous feed (F) solution. The optimal separation point must be balanced between maximising the yield and the separation. There are several levers available to optimise and control the extent of separation. These include the choice of extractant and manipulation of the reaction conditions such as pH, temperature, extractant concentration, and pre-neutralisation levels.
Selection of the extractant is a primary consideration in driving mass transfer across the interphase. For Co and Al, acidic extractants, facilitating a cation-exchange mechanism (as shown in Equation 3), have yielded superior selectivity (Makanyire et al., 2016). From these, CYANEX 272 was chosen due to its widespread availability and competitive separation efficiencies (Tait, 1993).
[3]
exploration thereof in this study is crucial due to the current unique feed composition. Furthermore, the effect of additional variables, i.e., the feed metal concentration was evaluated. To further deepen the understanding of variable effects on the separation efficiency, the use of OLIsoftware was included to identify the metal speciation under various conditions, which has not previously been explored in Al and Co separation studies. These results can be used to aid with experimental planning by means of identifying pH regions where the desired metal species will exist for complexation with CYANEX272. The OLI data can also be used as one layer of validation of the shake-out results.
In summary, the aim for this research was to optimally separate Co and Al from spent FTS catalyst by evaluating variables commonly studied in literature, ssuch as temperature, pH, pre-neutralisation, and additional variables such as feed metal concentration. The study furthermore used the OLI database to form part of both the experimental planning and to aid with experimental validation. After the optimal conditions for separation were determined, the secondary objective was to develop a conceptual design for a mixer-settler (typically used in SX) with specific focus on sizing the settler, as the need for the design was to obtain a high-level view of footprint required to be able to compare with pertraction technology (outside of the scope of this article).
Materials and methods
Metal speciation
OLI Studio software (version 11.5.1) was used to determine experimental ranges. According to the complexing mechanism shown in Equation 3, both Al (as Al3+) and Co (as Co2+) should bind with the phosphinic ions present in CYANEX 272 with active component bis(2,2,4-trimethylpentyl) phosphinic acid. The obtained speciation data of the metal ions at different aqueous pH and temperature conditions were used to guide the pH and temperature range at which the experiments should be conducted. The OLI databases were leveraged to create three sets of Pourbaix diagrams of Al and Co, as shown in Table 1.
Extraction kinetics
where Mn+ refers to a metal ion of n valence, (aq) to the aqueous phase, and (org) to the organic phase.
In terms of the variables affecting the separation efficiency, various studies have highlighted the importance of the aqueous solution pH (Tsakiridis, Agatzini-Leonardou, 2005; Torkaman et al., 2017). Other measurable variables include temperature, extractant concentration, and the extent of pre-neutralisation. Although the effects of these variables have been studied to some extent, the
The first experimental step was to quantify the minimum contact time necessary to reach equilibrium. For this, an aqueous phase containing 0.9 g/L Al and 0.1 g/L Co in deionised water (Millipore Milli-Q Plus Q-pack CPMQ004R1) was contacted for 2 h with a 20 vol.% CYANEX 272, 5 vol.% tributyl phosphate and the remainder made up of ShellSol 2325. Incremental samples were taken, separated, and analysed using inductively coupled plasma optical emission spectroscopy (ICP-OES; Agilent 5110). Additional details on the specifications and suppliers of the aforementioned reagents are provided in Table 2.
Table 1
OLI Studio Pourbaix diagram set-up conditions
Separation and recovery of cobalt and aluminium from spent gas-to-liquid catalysts
Table 2
Specifications of reagents used to prepare the synthetic aqueous and organic solutions
Reagent Purity
Cobalt(II) sulfate (CoSO4.7H2O) > 90%
Aluminium sulfate (Al2(SO4)3.16H2O) > 90%
CYANEX 272 > 85%
Supplier
Anyang General Chemical Co., Ltd
Anyang General Chemical Co., Ltd
Syensqo
ShellSol 2325 – CHEMQUEST AFRICA
Tributyl phosphate (TBP) > 98%
Ammonium hydroxide > 25%
Sodium hydroxide > 98%
Sulfuric acid > 98%
Extraction
Contact solution preparation
A synthetic aqueous solution was prepared to obtain the desired 9:1 Al:Co (0.9 g/L Al and 0.1 g/L Co) catalyst composition, based on literature and the analysed pregnant leach solution (PLS). After preparation, the solution was allowed to stand for approximately 2 hours to obtain a homogenous distribution. The concentration of the metals in the solution was confirmed using ICP-OES. The organic phase consisted of three components, namely the extractant, the diluent, and the modifier, which in this study were 20 vol.% CYANEX 272, 75 vol.% ShellSol 2325, and 5 vol.% TBP, respectively.
Effect of pre-neutralisation on extraction
To quantify the effect of pre-neutralisation on extraction, the organic and aqueous solutions were contacted at an O:A (organic to aqueous) ratio of 1:1 for 1 h at 40°C using a benchtop shaking incubator (Labotec Orbital Shaking Incubator, 200 rpm) to agitate the mixture. To determine the separation efficiency as a function of the extent of pre-neutralisation, 0.00–0.63 M NH4OH was added to the standard organic solution (20 vol.% CYANEX 272, 75 vol.% ShellSol 2325 and 5 vol.% TBP) before contacting with the aqueous phase. The NH4OH concentration range was based on the guidelines found in the CYANEX 272 technical datasheet, recommending a maximum of 50% addition of pre-neutralisation with respect to the extractant molar concentration (20 vol.% CYANEX 272, which is equivalent to 0.61 MCYANEX 272). After extraction, the mixture was poured into standard separation funnels to settle and separate.
Effect of aqueous feed pH on extraction
To determine the separation efficiency as a function of pH, the pH of the aqueous standard feed solution was adjusted using either H2SO4 or NaOH. Two different methods were used: (1) assessing the effect of pH on a solution containing one metal at a time because this is the usual method pursued in similar literature (Tsakiridis, Agatzini-Leonardou, 2005; Torkaman et al., 2017) and will allow for validation of experimental results, and (2) assessing the effect of pH on the standard feed solution containing both metals in a 9:1 mass-based Al:Co ratio (representative of the actual spent FTS solution). The pH values of the solutions were adjusted (with a calibrated Metrohm 744 pH – 6.0258.010 electrode) to
Fisher Scientific
Minema
LABCHEM
LABCHEM
obtain a pH range between ~1 and ~6 representing the effective zone (Tsakiridis, Agatzini-Leonardou, 2005) for both Al and Co extraction. The adjusted solutions were thereafter contacted with the standard organic phase (composition provided in the previous section) at 40°C, pre-neutralised to 50 mol.% CYANEX 272, shaken at 200 rpm for 1 h, separated, and analysed as described in the aforementioned.
Effect of temperature on extraction
The effect of temperature on the separation efficiency was conducted by contacting the solutions at an O:A ratio of 1:1 at varying temperatures (25–40°C). All organic solutions were preneutralised at 50 mol.% of CYANEX 272. All aqueous solutions were at a pH of 3.13.
Effect of feed metal concentration on extraction
The effect of feed metal concentration on extraction was determined by varying the metals in the feed from a concentration of 0–1 g/L, using the same reaction conditions as described in the aforementioned. The experiments were conducted by first keeping aluminium constant at 0.9 g/L and varying cobalt concentration whereafter cobalt concentration was kept constant at 0.1 g/L and aluminium concentration was varied.
Extraction isotherms
To obtain the extraction isotherms, the separation efficiency as a function of the feed concentration was calculated, which was varied from 0–10 g/L of Co or Al (as individual metals in solution). In addition to varying the Co and Al concentration in the aqueous solutions, different concentrations of the extractant in the organic phase (10–20 vol.% CYANEX 272) were used to ensure the mixture contained enough complexation sites to reach true equilibrium (note that TBP was kept constant at 5 vol.% and ShellSol 2325 was adjusted accordingly).
Scrubbing
During the scrubbing test work, the loaded organic phase obtained from the optimised extraction step was contacted with a scrub liquor. Two scrubbing methods were evaluated. According to Tsakiridis and Leonardou (2005), scrubbing with dilute sulfuric acid optimally removes co-extracted Co. Accordingly, in method one, scrub liquors with a pH ranging from 0.7–3.7 (0.1–0.0001 M H2SO4) were prepared and contacted with the loaded organic phase.
Separation and recovery of cobalt and aluminium from spent gas-to-liquid catalysts using CYANEX®
In method two, a metal-rich scrubbing liquor containing ~ 50 g/L Al (maximum solubility at room temperature) was prepared at the optimal pH (2.8) and contacted with the loaded organic phase (Botelho et al., 2019; Tsakiridis, Agatzini-Leonardou, 2005). In both cases, the phases were contacted in an O:A ratio of 4:1 for 1 hour at 40°C in a shaking incubator.
Stripping
During the stripping test work, the optimal scrubbed organic phase was mixed with an aqueous strip liquor. A strip liquor was prepared over a concentration range of 0.01–3.0 M H2SO4 (Botelho et al., 2019). An O:A ratio of 4:1 was used when contacting the two phases for 1 hour at 40°C in a shaking incubator.
Mathematical design of conventional mixer-settler
From the shake-out tests, the optimal separation conditions were determined, which provided enough information to design an industrial mixer-settler. The objective of the mixer-settler design was to obtain the relative size required to reach the determined equilibrium of the extraction phase for a given throughput. The performance of a mixer-settler is greatly determined by effective contact during mixing and the efficiency of the gravity separation step. Hence, for the simplified mixer-settler design, the two categories were modelled separately: firstly, the stages required for agitation (mixer) and, secondly, the area necessary for complete physical separation during the settling step. The McCabe-Thiele method was used to calculate the number of mixer stages required and the guidelines provided by Rydberg et al. (2004) were followed for the sizing of the settler.
After determining the number of stages, the dimensions (diameter/width) of the tank were varied until a design was obtained that met the five requirements presented in Table 3. According to the rules of thumb pointed out by Rydberg et al. (2004), the following tank dimension assumptions were made: i) height = width and ii) length = 4 × width. The final design geometry was adjusted to include a 20% overdesign factor to account for these assumptions. Once these requirements had been met, the entrainment layer was minimised to allow for complete separation of the two phases.
Results
Metal speciation
Figures 1(a) and (b) provide the Pourbaix diagrams for Al and Co, respectively, in the presence of sulfates.
Table 3
Requirements for a settler design to ensure gravity separation takes place
Requirement number
Requirement*
1 Qc/AI < vd
2 Qd/AI < vc
3 td > 2 min
4 Tav > 2 × T minD
5 Reynolds (Re) < 5000
*Qc and Qd are the overflow rates of the continuous and the dispersed phases (m3/s), AI refers to the area of the interface (m2), vc and vd refer to the terminal settling velocity of a droplet in the continuous and dispersed phase, respectively (m/s), td refers to the time required to cross the dispersed band (s), Tav and TminD are the average residence time of the dispersed phase(s) and the minimum residence time, respectively, as determined by Stokes’ law (s)
According to the cation-exchange mechanism provided in Equation 3, typical cations (like Al3+ and Co2+) should be the desired metal species that can complex with CYANEX 272, establishing separation. Pourbaix diagrams were used to obtain the pH ranges for the various species that could occur in such a feed system. Temperatures were varied but no significant change in stable regions was noted. The immunity region (where the pure metal will not dissolve and form ion complexes with the aqueous substance) of Al and Co is illustrated by the dark grey area in Figure 1. The black dashed lines (a and b) indicate the water reduction and oxidation lines, respectively. According to Figure 1(a), the desired species of Al should exist as AlSO4+ at pH between 0–3.4, while according to Figure 1(b), Co2+ should exist at pH between 0–7.8. The desired complexing species of both Co and Al are expected to be present at low pH values, hence competition between the metals to bind with CYANEX 272 can be expected at pH between 0–3.8. Neither the Al nor Co desired complexing species are predicted to exist at high pH values (> 7.8), where precipitation would likely occur and therefore experiments above this region are not recommended.
Extraction kinetics
As shown in Figure 2, equilibrium was reached after ~15 min of contact time. This is comparable with literature when using similar conditions (Torkaman et al., 2017). Although 15 minutes were
Figure 1—Pourbaix diagrams of (a) 0.9 g/L aluminium and (b) 0.1 g/L cobalt at 25°C, 1 atm in sulfate solutions
Separation and recovery of cobalt and aluminium from spent gas-to-liquid catalysts using
adequate under these conditions, all remaining experiments were conducted for 1 hour to ensure that equilibrium was reached when using varying reaction conditions.
Extraction
Effect of preneutralisation on extraction
The effect of preneutralisation on extraction as a function of the NH4OH concentration is shown in Figure 3. Firstly, without preneutralisation, poor Al extraction (26%) and separation efficiency (15%) were observed. This is to be expected, because in the absence of preneutralisation, i.e., no pH control, H+ will be exchanged from the extractant to the aqueous phases, resulting in a suboptimal equilibrium pH and a decrease in extraction performance. Preneutralisation with NH4OH resulted in less variability in pH, resulting in a significant improvement in Al extraction and separation efficiency. According to Figure 3, the optimal point of separation was attained when preneutralising with 0.38 M NH4OH (~60 mol.% of CYANEX 272), yielding 98% Al extraction with a separation efficiency of 83%. However, the CYANEX 272 technical datasheet recommends that no more than 50 mol.% of the extractant should be preneutralised to avoid the risk of sacrificing selectivity, the remainder of the work was conducted at the recommended 50 mol.% (dashed line), which resulted in 87% extraction of aluminium and 17% extraction of cobalt.
Effect of aqueous feed pH on extraction
The dependence of extraction efficiency on pH is shown for the mixed and single metal feeds in Figure 4(a) and Figure 4(b), respectively.
A similar trend in experimental results, as illustrated in Figure 4(a), was observed by Tsakiridis and Leonardou (2005). In both cases, Al was first extracted to the organic phase in a pH range of 1 to 4, before reaching a plateau above pH 4. According to Tsakiridis and Leonardou (2005), 98.7% Al was extracted at pH ~ 2.6, while only 12% Co was extracted at pH 3.15. Accordingly, the optimal separation was between pH 2.5 and 3.5, which also compares with results obtained by Suzuki et al. (2012) who found an optimal extraction pH for Co and Al of 2.8. The difference in the optimal pH values between the data presented in literature and current data (optimum selectivity at pH 3.13), is due to the difference in feed compositions as well as experimental designs. Suzuki et al. (2012) and Tsakiridis and Leonardou (2005) individually determined the
Figure 2—0.9 g/L aluminium solvent extraction kinetics at 40°C and 1 atm with 50 mol.% preneutralisation of 20 vol.% CYANEX 272, 75 vol.% ShellSol 2325 and 5 vol.% TBP over 2 hours
Figure 3—Effect of preneutralisation concentration on 0.1 g/L cobalt and 0.9 g/L aluminium extraction at pH 3.1, 40 °C and 1 atm with 50 mol.% preneutralisation of 20 vol.% CYANEX 272, 75 vol.% ShellSol 2325, and 5 vol.% TBP over 1 hour
pH dependence of extraction of each metal. In Figure 4(b), both the literature and the current data for individual metal extractions are shown, whereby the experimental data obtained differed with a mean absolute error of 4% on both the Al and Co trends. When a single metal aqueous feed is contacted with an organic phase, no competition amongst metals can be evaluated and therefore cannot truly represent mixture selectivity. When comparing Figure 4(a) and Figure 4(b), the competition of Co and Al ions can be noted, which indicates the internal metal competition for complexation with CYANEX 272, as predicted from the OLI database, shown in Figures 1(a) and (b). At pH 2, for example, more than 90% Al was extracted according to Figure 4(b). However, when Al competes with Co, as shown in Figure 4(a), less than 60% of Al was extracted with ~ 30% co-extraction of Co at pH 2. These results confirm the importance of conducting extraction experiments on
Figure 4—Extraction efficiency dependence on aqueous feed solution pH when the experiment is conducted on (a) the standard feed solution (0.1 g/L cobalt and 0.9 g/L aluminium) and (b) a single metal (0.1 g/L cobalt and separately 0.9 g/L aluminium) in solution: dots = experimental data points and lines = data obtained from literature (Tsakiridis, Agatzini-Leonardou, 2005). Conducted at 40°C and 1 atm with 50 mol.% preneutralisation of 20 vol.% CYANEX 272, 75 vol.% ShellSol 2325, and 5 vol.% TBP over 1 hour
Separation and recovery of cobalt and aluminium from spent gas-to-liquid catalysts using CYANEX® 272
representative mixtures to capture underlying metal competition. When combining the stability diagrams generated in Figures 1(a) and (b) with the extraction results presented in Figures 4(a) and (b), it seems that the Al that was extracted in the pH range 0–3.4 might have been in the form of AlSO4+ rather than the suggested Al3+ form, according to the mechanism shown in Equation 3. It is suggested that additional tests be conducted to further understand the desired metal species that may be required. Furthermore, it is noteworthy to mention that an overall mean absolute experimental error of 5% was obtained.
Effect of temperature on extraction
Figure 5 shows the effect of temperature on the overall extraction efficiency. The temperature only affected the Al extraction within the range of 25°C–40°C, with no significant effect above 40°C. These results compare well with literature findings (Torkaman et al., 2017). Therefore, the temperature of all the experiments was set at 40°C, where optimal separation, with a minimum energy input, is expected.
Effect of feed metal concentration on extraction
The distribution coefficients of Co and Al at varying feed metal concentrations are provided in Figure 6. The data demonstrate that a higher feed metal concentration results in a lower distribution coefficient, which is expected due to factors such as competitive binding (Kislik, 2012). The extraction efficiency is determined by the distribution coefficient of the metal and is therefore also a function of the metal concentration.
Keeping in mind that all the experiments were conducted at pH 3.13 and 50 mol.% preneutralisation, the selectivity changed as the feed metal concentration changed. As a result, in an industry where the feed metal concentration may vary, it is important to adjust the contact pH to ensure optimal conditions.
Extraction isotherms
Figure 7(a) and Figure 7(b) show the extraction isotherms of Al and Co, respectively, varying in concentrations from 0-1 10 g/L as induvial metals in solution, as specified in the Materials and methods section. While the amount of Al extraction increased with increasing CYANEX 272 concentration, the effect was less significant than the effect observed when extracting Co, as can be observed in Figure 7(b). More Al is transferred to the organic phase when comparing the two metals according to similar concentrations. Unlike Co distribution, Al has a strong drive through the whole concentration range. The extractant is also more selective towards Al, which correlates with the higher distribution coefficient obtained (Figure 6).
Scrubbing
Two scrubbing methods were evaluated: i) dilute sulfuric acid at pH 0.7–3.7 (0.1–0.0001 M H2SO4) and ii) an Al-enriched scrubbing liquor (50 g/L) at the optimal pH of 2.8. In both cases, the aim was to scrub off the Co while leaving Al in the organic phase. For the scrubbing, a loaded organic phase (consisting of 0.78 g/L Al and 0.017 g/L of Co) was obtained from extraction at pH 3.13, shaken at 200 rpm, at 40 °C, and preneutralised at 50 mol.% CYANEX 272 (20 vol.%).
Method 1
The scrubbing efficiency from method one is shown in Figure 8. From Figure 8, which illustrates the scrubbing efficiency as a function of scrubbing liquor pH, it is apparent that using diluted
Figure 5—Extraction efficiency dependence on temperature. Conducted at an aqueous solution (containing 0.1 g/L cobalt and 0.9 g/L aluminium) pH 3.13 and 1 atm with 50 mol.% preneutralisation of 20 vol.% CYANEX 272, 75 vol.% ShellSol 2325, and 5 vol.% TBP over 1 hour
Figure 6—Distribution coefficients of cobalt and aluminium at varying feed metal concentrations. Conducted at an aqueous solution pH 3.13 and 1 atm with 50 mol.% preneutralisation of 20 vol.% CYANEX 272, 75 vol.% ShellSol 2325, and 5 vol.% TBP over 1 hour
Figure 7—(a) Aluminium and (b) cobalt extraction isotherms with 1020 vol.% CYANEX 272. Conducted at 40 °C and 1 atm with 50 mol.% preneutralisation of 10-20 vol.% CYANEX 272,75-85 vol.% ShellSol 2325, and 5 vol.% TBP over 1 hour
H2SO4 as a scrub liquor was not selective towards Co or Al. Hence method one is not suitable for this application.
Method 2
In this single experiment, 50 g/L Al was added to the scrubbing liquor which was set to pH 2.8. Most (99.7%) of the Co was removed from the loaded organic phase (originally containing 0.017 g/L of Co and 0.78 g/L of Al), resulting in a Co concentration of 0.05 mg/L and an Al concentration of 1.13 g/L due to replacement with Al. At this point, the Co concentration is 44 ppm
Separation and recovery of cobalt and aluminium from spent gas-to-liquid catalysts using
Figure 8—Scrubbing efficiency with H2SO4 over a range of pH values. Conducted at 25°C and 1 atm with the loaded organic phase consisting of 0.78 g/L aluminium, 0.017 g/L cobalt in 50 mol.%, preneutralised CYANEX 272 (20 vol.%), 75 vol.% ShellSol 2325, and 5 vol.% TBP over 1 hour
(with respect to Al concentration) meeting the <100 ppm of Co target set in the Introduction section. Method two was therefore the scrubbing method of choice.
Stripping
To optimise the Al transfer from the organic phase to the aqueous strip liquor, the sulfuric acid concentration was varied from 0.01 M–3.0 M. The resulting stripping efficiency is shown in Figure 9. The point of optimal stripping seems to be at a concentration of 1.0 M sulfuric acid, (which results in an ~80% stripping efficiency of Al and 99.3% stripping efficiency of Co). The Final loaded strip solution therefore contains 0.91 g/L of Al and 5.4 x 10-5 g/L of Co resulting in a final strip solution of ~54 ppm Co relative to Al while recovering ~70% of the original Al in the feed solution, thus meeting the initial target of <100 ppm of Co.
Mixer-settler design
The number of stages required to obtain the extraction determined by the shake-out tests was obtained with the McCabe-Thiele method as shown in Figure 10, (which provides a section of the isotherm provided in Figure 7 that is relevant to this application with lower concentrations). Two mixing stages were deemed necessary to attain an extraction efficiency of 87%, as demonstrated in Figure 4(a). Furthermore, the minimum decanter size necessary to ensure efficient gravity separation was determined to be at 1.5 m high, 1.5 m wide, and 6 m in length. The main results of the settler size calculations are presented in Table 4. Only a settler was sized, as this is generally the largest equipment compared (Van Nguyen et al., 2023) to the mixer and the sizing was required to conduct a high-level footprint comparison with pertraction based extraction (outside of the scope of this article).
The overflow rate was less than the settling velocity, ensuring that settling occurred before extraction of a phase at the overflow (applying to both the dispersed and continuous phase). The coalescence time exceeded 2 minutes, ensuring enough time for a molecule to cross the dispersion phase. Finally, the Reynolds number < 5000 ensured that turbulence should not hinder the separation.
Conclusions
The investigation aimed to optimise the separation process for a PLS with an Al:Co ratio of 9:1. Optimal conditions at 40 °C, 1 h contact time, with an aqueous feed solution pH of 3.13 when using CYANEX® 272 as extractant was established. Notably, scrubbing with diluted sulfuric acid lacked efficiency, while using 50 g/L Al at
Figure 9—Effect of sulfuric acid concentration on the stripping efficiency of aluminium. Conducted at 25°C and 1 atm on the scrubbed organic containing 1.13 g/L aluminium and 0.05 mg/L Co in 50 mol.% preneutralised CYANEX 272 (20 vol.%), 75 vol.% ShellSol 2325, and 5 vol.% TBP over 1 hour
Figure 10—McCabe-Thiele plot determining mixer stages required
Table 4
Mixer-settler design results
Requirement
Requirement 1: Qc
Requirement 2: QD <vc
Requirement 3: td > 2 min
min Requirement 4: tav > 2 × tminD
pH 2.8 proved effective. Optimal stripping was attained when using an aqueous solution with 1 M H2SO4 and 1 hour contact time. This process obtained 54 ppm Co in the loaded strip liquor, reaching the desired separation target of achieving less than <100 ppm of Co. The predicted coexistence of complexing Al and Co species at
Separation and recovery of cobalt and aluminium from spent gas-to-liquid catalysts using CYANEX® 272
pH < 4 led to competitive behaviour between the two metals when in contact with CYANEX® 272. This phenomenon was evident in the exchange of metal ions, notably Co, observed at lower pH levels. Moreover, the extraction efficiency proved to be influenced not only by pH and preneutralisation but also by the feed metal concentration. Industries should consider varying the contact pH in response to different feed metal concentrations to optimise overall yield. According to the mixer-settler design study, two mixer stages was required. Additionally, a settler sizing of 1.5 m in height, 1.5 m in width, and 6 m in length was recommended to ensure adequate residence time for gravity separation, catering for a flow rate of 10 m3/h. Moving forward, potential areas for further investigation could include validating the identified Al species in the presence of sulfates, exploring the impact of other potential variables on extraction efficiency, and conducting pilot-scale tests to validate the scalability of the proposed conditions. Additional experimental runs are also recommended to obtain a better representation of the isotherm shapes.
CRediT statement
MK: Investigation, validation, formal analysis, writing - original draft; HMK: Conceptualisation, methodology, writing - review and editing; DvdW: Conceptualisation, methodology, writing - review and editing
References
Botelho, A., Dreisinger, D., Espinosa, D. 2019. A review of nickel, copper and cobalt recovery by chelating ion exchange resins from mining processes and mining tailings. Mining, Metallurgy and Exploration, vol. 36, no. 1, pp. 199–213. http://dx.doi.org/10.1007/s42461-018-0016-8
Chong, S., Hawker, W., Vaughan, J. 2013. Selective reductive leaching of oxidised cobalt containing residue. Minerals Engineering, vol. 54, pp. 82–87. http://dx.doi.org/10.1016/j.mineng.2013.04.004
Dhiman, S., Gupta, B. 2019. Partition studies on cobalt and recycling of valuable metals from waste Li-ion batteries via solvent extraction and chemical precipitation. Journal of Cleaner Production, vol. 225, pp. 820–832. https://doi.org/10.1016/j.jclepro.2019.04.004
Ferreira, D.A., Prados, L.M.Z., Majuste, D., Mansur, M.B. 2009. Hydrometallurgical separation of aluminium, cobalt, copper and lithium from spent Li-ion batteries. Journal of Power Sources, vol. 187, pp. 238–246. http://dx.doi.org/10.1016/j.jpowsour.2008.10.077
International Plant Nutrition Institute, 2014. Nutri-facts: Agronomic fact sheet on crop nutrients. http://www.ipni.net/nutrifacts-northamerican [accessed 02 September 2024]
Jacobs, G., Tapan, K.D., Yongqing, Z., Jinlin, L. 2002. FischerTropsch synthesis: support loading and promotor effects on the reducibility of cobalt catalysts. Applied Catalysis A: General, vol. 233, pp. 263–281. http://dx.doi.org/10.1016/S0926-860X(02)00195-3
Jahangiri, H., Bennett, J., Mahjoubi, P. 2014. A review of advanced catalyst development for Fischer-Tropsch synthesis of hydrocarbons from biomass derived syngas. Catalysis Science and Technology, vol. 4, no. 8, pp. 2210–2229. http://dx.doi.org/10.1039/C4CY00327F
Kislik, V.S. 2012. Competitive Complexation/Solvation Theory of Solvent Extraction. 1st ed. Elsevier/Amsterdam.
http://dx.doi.org/10.1016/B978-0-444-53778-2.10007-X
Liang, K., Zhang, C., Xiang, H., Yang, Y. 2019. Effects of modified SiO2 on H2 and CO adsorption and hydrogenation of ironbased catalysts. Journal of Fuel Chemistry and Technology, vol. 47, no. 7, pp. 769–779.
http://dx.doi.org/10.1016/S1872-5813(19)30033-7
Makanyire, T., Sanchez-Segado, S., Jha, A. 2016. Separation and recovery of critical metal ions using ionic liquids. Advanced Manufacturing, vol. 4, pp. 33–46.
http://dx.doi.org/10.1007/s40436-015-0132-3
Rydberg, A., Cox, M., Musikas, C., Choppin, G.R. 2004. Solvent Extraction Principles and Practice. 2nd ed. CRC Press, Boca Raton. https://doi.org/10.1201/9780203021460
Sole, K. 2008. Solvent extraction in the hydrometallurgical processing and purification of metals: process design and selected applications. Solvent Extraction and Liquid Membranes: Fundamentals and Applications in New Materials. M Aguilar and J.L Cortina (Eds). 1st ed. CRC Press, Boca Raton. pp. 141–200.
Suzuki, T., Nakamura, T., Inoue, Y. 2012. A hydrometallurgical process for the separation of aluminium, cobalt, copper and lithium in acidic sulfate media. Seperation and Purification Technology, vol. 98, pp. 396–401. http://dx.doi.org/10.1016/j.seppur.2012.06.034
Tait, B.K. 1993. Cobalt-nickel separation: the extraction of cobalt(II) and nickel(II) by Cyanex 301, Cyanex 302 and Cyanex 272. Hydrometallurgy, vol. 32, no. 3, pp. 365–372. https://doi.org/10.1016/0304-386X%2893%2990047-H
Torkaman, R., Asadollahzadeh, M., Torab-Mostaedi, M., Maragheh, M. G. 2017. Recovery of cobalt from spent lithium-ion batteries by using acidic and basic extractants in solvent extraction process. Separation and Purification Technology, vol. 186, pp. 318–325. https://doi.org/10.1016/j.seppur.2017.06. 023
Tsakiridis, P., Agatzini-Leonardou, S. 2005. Solvent extraction of aluminium in the presence of cobalt, nickel and magnesium from sulphate solutions by Cyanex 272. Hydrometallurgy, vol. 80, pp. 90–97. http://dx.doi.org/10.1016/j. hydromet.2005.07.002
Van Nguyen, T., Van Nguyen, D., Ngo, Q.H., Nguyen, T.L. 2023. A Calculation, design and manufacture of system box-type mixersettler used in rare earth solvent extraction process at laboratory scale. Nuclear Science and Technology, vol. 12, pp. 20–27. http://dx.doi.org/10.53747/nst.v12i2.346 u
Affiliation:
Advanced Materials Division, Mintek, Randburg, Johannesburg
Correspondence to: C. Felix
Email: CecilF@mintek.co.za
Dates:
Received: 20 Mar. 2025
Revised: 25 Mar. 2025
Accepted: 25 Mar. 2025
Published: May 2025
How to cite:
Felix, C., Barron, O. 2025. Roll-to-roll coating methods for the manufacture of polymer electrolyte membrane fuel cell membrane electrode assemblies. Journal of the Southern African Institute of Mining and Metallurgy, vol. 125, no. 5, pp. 267–272
DOI ID:
https://doi.org/10.17159/2411-9717/818/2025
ORCiD:
C. Felix
http://orcid.org/0000-0002-0293-9682
O. Barron
http://orcid.org/0000-0003-2708-0533
This paper is based on a presentation given at the 2ND Southern African Hydrogen and Fuel Cell Conference 2025, 7-8 April 2025, Southern Sun Rosebank, Johannesburg
Roll-to-roll coating methods for the manufacture of polymer electrolyte membrane fuel cell membrane electrode assemblies
by C. Felix, O. Barron
Abstract
In 2023, polymer electrolyte membrane fuel cell membrane electrode assemblies accounted for about 62% of the global fuel cell market share and this demand is expected to grow. Commercialisation requires technologies capable of cost-effective, high-volume manufacturing. The membrane electrode assembly is crucial and represents a significant cost to polymer electrolyte membrane fuel cell manufacturing. Manufacturing the electrode components for membrane electrode assemblies at high volumes can reduce production costs while delivering on the forecasted industry demands. Roll-to-roll is a mature coating technology applied in various manufacturing industries such as batteries, printed electronics, microelectronics, photovoltaics, printing, paper consumables, etc., with significant potential for polymer electrolyte membrane fuel cell manufacturing. While the high-volume production process for membrane electrode assembly manufacturing differs from laboratory-scale, the membrane electrode assembly structure remains unchanged. Roll-to-roll is a continuous coating process with speeds of 10 m/min being achieved. In this work, MicrogravureTM and slot-die coating methods have been adopted and have shown promising outcomes for producing gas diffusion electrodes for membrane electrode assemblies. An overview of the MicrogravureTM and slotdie coating methodologies is provided, and a discussion on preliminary results and challenges are presented. Work on MicrogravureTM gas diffusion electrode manufacturing has already achieved comparable performance with spray-coated gas diffusion electrodes. The influence of coating parameters and drying on the coated catalyst layer quality is discussed.
Keywords
polymer electrolyte membrane fuel cell, gas diffusion electrode, slot-die, MicrogravureTM, rollto-roll, coating window
Introduction
Fuel cells (FC) are electrochemical devices that directly and efficiently convert chemical energy into electrical energy. Polymer electrolyte membrane fuel cells (PEMFC) especially, are a crucial step towards zero carbon emissions energy utilisation and as of 2023, accounted for about 62% of the fuel cell market share (Grand View Research, 2023). However, issues related to PEMFC cost and performance are key issues that still need solving for commercialisation to be realised (Liu et al., 2023). The use of platinum (Pt) catalysts still represents 55% of the cost of PEMFC manufacturing (Liu et al., 2023), however, until alternative non-platinum group metal catalysts with performance parity to Pt catalysts are developed, alternative manufacturing processes for reducing the membrane electrode assembly (MEA) cost should be explored. One approach involves the MEA manufacturing process itself. Various methods exist to coat the MEA catalyst layers (CLs). These include paint brushing, spray-coating, doctor blade coating, gravure coating, ink-jet printing, slot-die coating, and electrospinning (Medina et al., 2023). Spray-coating is typically utilised at laboratory and small-scale production levels to coat the CL onto either the gas diffusion layer (GDL) to form the MEA or directly onto the membrane to form catalyst-coated membrane (CCM) MEAs. Although CCMs are desired due to the improved MEA performances, many challenges exist due to the swelling of the membrane when contacted with the solvents in the catalyst ink. Spray-coating is time-consuming since multiple thin layers are coated to build up the CL and require many spray coaters and operators to meet the demands of the emerging fuel cell industry. An analysis reported by Mark Debe (2012) revealed that by 2030, 10% of vehicles will be FC-powered and meeting the demand of 15 million FC vehicles per annum (each with a stack comprising 300 MEAs), production speeds of 20 m2/min will be required. Medina and others estimated that spray coaters can coat at 0.2 m2/min with one nozzle, thus facilities with 100 coating nozzles will be required to meet the 20 m2/min demand (Medina et al., 2023). Continuous production processes using roll-to-roll (R2R)
Roll-to-roll coating methods for the manufacture of polymer electrolyte membrane fuel cell
technologies are more suitable for meeting production targets. R2R refers to a family of manufacturing techniques in which a flexible substrate is continuously coated as it is unwound from a stock roll and transferred to a rewinding roll (Rashid et al., 2021). R2R manufacturing offers significant cost reductions and although initial capital costs can be high, these costs can often be recovered through economies of scale. R2R methods can be either self-metered or premetered, depending on how fluid metering is achieved. Metering is the process of controlling the rate at which fluids are applied during coating. Examples of self-metered R2R methods are gravure coating, knife, comma bar, etc. Pre-metered methods include slot-die, spray coating, extrusion die, slide curtain, etc. This work presents results from an R2R MicrogravureTM method (Yasui-Seiki Mini-Labo Deluxe, Japan) and a benchtop slot-die coating method (Ossila, United Kingdom). Both techniques are suitable for highvolume production, however, various factors influence the quality of the coatings and may limit the Pt loadings required for specific PEMFC applications. Due to the many variables that need to be considered for slot-die, significant work has been dedicated to labscale optimisation and reported here. The challenges experienced during the practical operation of these methods are presented and discussed.
MicrogravureTM overview
Gravure cylinder coating originated in the printing industry and is desirable for producing thin, uniform films at high speeds. The gravure cylinder, also called gravure roller, can typically operate in forward and reverse modes. The surface of the gravure cylinder is engraved or etched with a discrete pattern of cells that transfer the fluid from a tray to the substrate. The gravure cylinders can be engraved with various cell pattern designs such as quadrangular, tri-helical, and cells linked by small channels (Hewson et al., 2011). Figure 1 illustrates a simplified overview of the gravure coating process. Gravure coating involves the patterned cylinder partially submerged in a tray with the fluid. The gravure cylinder is rotated in the fluid and the engraved cells are filled. The doctor blade trims excess fluid from the gravure cylinder. The substrate is tensioned over a series of rollers, known as a web, and moves over the gravure cylinder. Fluid is transferred to the substrate where contact occurs, forming the coating. The cylinder design can affect the coating thickness since the grooves affect the cell opening per unit area, the average volume (cm3/m2), and the line count (lines per inch) (Kapur et al., 2011). MicrogravureTM is a variation of gravure coating that uses a smaller diameter cylinder than regular gravure coating. The smaller diameter cylinder provides a smaller contact area and allows for a more stable coating bead to form. A stable bead is crucial for forming uniform thin coatings. MicrogravureTM is a reverse kiss gravure coating method. The ‘kiss’ indicates that the gravure cylinder coats the substrate through a ‘kissing’ action and there is no backing roll to trap the substrate against the cylinder. During coating, the engraved cylinder’s rotational direction is opposite to the web direction. Thus, the coating is applied to the substrate in a shearing manner (Yasui-Seiki, n.d.).
MicrogravureTM coating parameters
MicrogravureTM is a self-metered method and the coating thickness depends on the gravure cylinder design, cell pick-out, cylinder speed, and web speed (the substrate's speed). Cell pick-out refers to the ink transfer process from the gravure cylinder cells to the substrate, with each cylinder design having different cell pick-out rates. MicrogravureTM cylinders also come with varied cell volumes
similar to gravure coating. Poor pick-out leads to macroscopically defective coatings. The pick-out can be fine-tuned by precise control of the speed ratio, which is defined as the gravure cylinder speed (m/min)/web speed (m/min) (Liu et al., 2023; Park et al., 2020; Mauger et al., 2018). During coating, the fluid should be replenished at a rate equal to the consumption rate to prevent the cylinder from running dry or the fluid tray from overflowing. Both instances can affect the quality of the coating. Other parameters that affect the coating are the fluid’s rheological properties (viscosity, shear thinning, shear thickening, etc.) and wettability. Fluids with contact angles < 90° are generally considered to have good wettability while fluids with contact angles > 90° have poor wettability and will be difficult to coat or result in defects. However, for coating on PEMFC gas diffusion layers (GDLs), fluids with very low contact angles may lead to excessive penetration, which may restrict gas transport through the GDL as well as impact the balance of proton and electron conductivity, leading to decreased MEA performance (Ding, Harris, 2017). Figure 2 shows photomicrographs of defective coatings obtained when (a) the catalyst ink formulation and coater web speed are not optimum and (b) the coater web speed and gravure roller speed are not optimum. In Figure 2 (a), the catalyst ink’s contact angle was too high to sufficiently wet and penetrate the GDL’s microporous layer, resulting in a smudged coating with an average Pt loading of 0.01 mg/cm2 Slower web speeds may have resulted in better coating outcomes for this catalyst ink; however, slower speeds are not beneficial at production scales. A solution would be to reduce the catalyst ink’s water content to lower the contact angle on the GDL’s hydrophobic surface; increasing the alcohol content reduces the catalyst ink’s contact angle on the GDL surface. The GDL hydrophobicity is also an important property that needs to be considered when formulating the catalyst ink. In Figure 2 (b), the coater web speed and gravure roller speed were mismatched, resulting in poor cell pick-out and a streaked coating.
MicrogravureTM limitations and challenges
MicrogravureTM was designed to achieve ultra-thin uniform coatings. Wet coating thicknesses between 0.8 to 80 μm can be achieved using cylinders with different engraved cell volumes. Mauger et al. (2018) used MicrogravureTM to produce CLs in PEMFCs. They used a dilute catalyst ink of 3.2 wt.% catalyst and achieved Pt loadings between 0.1 - 0.13 mg/cm2 by varying the speed ratio. Although these Pt loadings are United States Department of Energy targets for light-duty vehicle applications (U.S Department of Energy, 2023), it is insufficient for heavy-duty applications where PEMFCs will play a key role. For heavy-duty vehicle applications, Pt loadings > 0.3 mg/cm2 are still required to meet the long durability of 30,000 hours or 1 million miles carrying heavy load (Sharma, et al., 2022). To explore the Pt loading limits of the MicrogravureTM method, an 8.3 wt.% catalyst ink was coated
Figure 1—Simplified illustration of the Gravure/MicrogravureTM coating process
Roll-to-roll
with varying speed ratios. XRF Pt loadings of 0.164 - 0.29 mg/cm2 were obtained when the speed ratio increased from 0.94–3.14. The Pt loading increased as the speed ratio increased until it reached a plateau when the gravure cylinder reached its limits. Below a speed ratio of 0.94, defective CLs were obtained, while at a speed ratio of 2.5, the Pt loading plateaued, indicating that cell pick-out was at its maximum. Increasing the speed ratio beyond this point resulted in no further increase in Pt loading. A higher catalyst wt.% ink extended the coating range to Pt loadings close to the requirements for heavy-duty PEMFC applications. Higher Pt loadings may be achieved using gravure cylinders with higher cell volumes or catalyst inks with higher catalyst wt.%, however, these may come with additional coating challenges. Since MicrogravureTM is a continuous coating method, the CL is coated in a single pass compared to spray coating where multiple thin layers are coated to build the final CL. Coating the CL in a single pass results in a thick wet layer, prone to cracking during drying. Although researchers are divided on the role of cracks in the CL, eliminating cracks during single-pass CL coating remains a challenge that needs to be addressed for R2R coating methods. Figure 3 graphically illustrates how the Pt loading and CL surface crack percentage increased as the speed ratio increased. Surface cracking strongly correlates to the thickness of the catalyst layer where layers above 0.23 mgPt/cm2 experienced significant surface cracking.
Figure 4 shows photomicrographs of catalyst layers corresponding to the speed ratios illustrated in Figure 3. Surface cracking can be seen to increase with increasing Pt loading. When the wet catalyst layer thickness increases, the surface layer dries faster than the inner layer, trapping solvent molecules. The only way for the trapped solvent molecules to escape is to cause stress and break through the dried surface layer, causing cracks in the surface. Adding higher boiling point solvents and crack-inhibiting additives to the catalyst inks have been explored to reduce surface cracking (Liu et al., 2024; Hasegawa, et al., 2021).
Slot-die overview
The invention of slot-die is attributed to Albert E. Beguin (Sarka, Tobis, 2022). Slot-die coating is a one-dimensional technique where the substrate moves past the coating head and is coated (Vak, et al., 2016). Slot-die can be efficiently used to obtain coatings with high uniformity (Sharma et al., 2022; Creel, et al., 2022). Slot-die minimises waste by allowing precise control of the dispensed fluid. Slot-die is a pre-metered coating method consisting of a coating head and several components assembled such that the fluid enclosed in the head becomes pressurised. The fluid is forced through a slot in the head to produce uniform coatings (Vak et al., 2016). The fluid can be supplied to the head through a precision pump or compressed air. During coating, the fluid is pushed into the head
and the pressure inside the head forces the fluid through a slot and onto the tensioned moving substrate. The substrate moves past the slot-die head and the fluid is continuously coated onto the substrate. After solvent evaporation and solidification of the particles, a dry uniform film on the substrate can be obtained (Ding et al., 2016). Constant fluid flow is required to prevent coating defects, thus highprecision pumps or flow meters are required to control the fluid feed. Fluctuations in the fluid flow rate can be caused by variations in viscosity, air bubbles in the system, temperature changes, and
Figure 2—Defective coatings obtained with (a) a high contact angle catalyst ink and (b) incorrect speed ratio
Figure 3—MicrogravureTM Pt loading calibration and associated catalyst layer crack percentages
Figure 4—MicrogravureTM coatings on the Freudenberg H23C6 GDL: (a) uncoated GDL, (b) coated at speed ratio = 0.94, (c) coated at speed ratio = 2.20, (d) coated at speed ratio = 2.51
Roll-to-roll coating methods for the manufacture of polymer electrolyte membrane fuel cell
mechanical wear and tear of the flow control equipment, which will emanate as defective layers. Figure 5 shows a schematic of the slot-die coating process. Similar to gravure coating, slot-die can be coupled to roll-to-roll (R2R) processes for high-volume MEA production. Slot-die offers advantages over conventional and ultrasonic spray coating methods such as scalability, lower operational costs, and higher production rates (Sharma et al., 2022).
Slot-die coating parameters
During slot-die, the coating thickness is controlled by the gap between the slot-die-head lips and the substrate, the slot-die-lip gap, the web speed, and other factors related to the coating fluid (Ding et al., 2016; Xie et al., 2021). The relative speed between the slot-die head and the substrate translates to the coating speed (Xie, et al., 2021). The coating speed, gap height, ink rheology (viscosity, shear thinning, etc.), substrate properties (roughness, hydrophobicity, hydrophilicity, etc.), and drying conditions can influence the quality of the obtained coatings. If proper control of these conditions is not met, defects such as ribbing, barring, rivulets, and air entrainment can occur (Ding, Harris, 2017; Ding et al., 2016). Figure 6 shows examples of coating defects that occurred when the optimum coating conditions were not met. The wet coating thickness (twet) in traditional slot-die coating can vary from ten to a few hundred microns. With a tensioned web, a wet thickness of less than 5 microns can be obtained (Ding et al., 2016). Wet thickness can be calculated using Equation 1 where v is the coating speed (cm/s), Q is the fluid dispense rate (cm3/s) and w is the coating width (cm). Stähler et al. (2019) used a similar equation to determine the wet thickness. Determining twet is crucial for setting the gap height to prevent the gap from being too small, causing scraping and pooling of the coated layer or too large, affecting the formation of a stable coating bead and leading to defects.
Determining operating limits for slot-die is a complex process where various factors and competing forces are taken into account, as discussed elsewhere (Creel et al., 2022; Ding, et al., 2016). Figure 7 shows coating windows developed for three catalyst inks utilising the same 50 wt.% Pt/C catalyst but different ionomer-tocarbon (I/C) and water/alcohol ratios. The coating window is a stable operational window where uniform and defect-free coatings are possible (Chang et al., 2007). Contrary to Newtonian fluids, fluids such as colloidal suspensions and polymeric solutions can have very complex rheological properties such as shear thinning, extension thickening, and viscoelasticity, which can affect the force balance in the coating bead and the subsequent operation limit of the coating process (Ding et al., 2016). The catalyst ink for MEA manufacturing is a colloidal suspension consisting of particles and
ionomers, sometimes with polymer additives that typically exhibit shear thinning when shearing is applied. However, the rheological properties are strongly influenced by the composition of the ink, such as catalyst type (support type and metal %), ionomer content, and solvent type. A simple viscocapillary model as described by Creel et al. (2022), Koh et al. (2012), Higgens and Scriven (1980), and Lee et al. (2011) was adopted to develop the coating windows. The parameter inputs for this model were the viscosity of the ink (N/m2.s), surface tension (N/m), coating width (m), gap height (m), upstream lip width (m), downstream lip width (m), and static and dynamic contact angles. From Figure 7 it is observed that increasing the I/C ratio from 0.9 (black coating window) to 1.8 (red coating window) resulted in a slight broadening and position shifting of the coating window. Higher ink flow rates were required at the same coating speed for the 1.8 I/C ratio catalyst ink to obtain defect-free coatings. The coating windows for the catalyst ink with a water/alcohol ratio of 9 (green coating window) had a very narrow coating window due to the low wettability of this ink. The high water content of the catalyst ink caused a contact angle of 120° on the GDL. Although a narrow coating window was produced, Pt loadings in this coating window were limited to ~0.1 mg/cm2, a loading too restrictive for MEA manufacturing. It should be noted that coating windows are strongly influenced by the GDL type as the surface tension will change.
Slot-die limitations and challenges
Slot-die coating equipment generally requires a higher upfront investment cost than simpler coating methods. Slot-die coating is also more complex and requires precise control of coating variables, which are often obtained through coating window development. Fluids have different coating windows depending on the fluid makeup. Colloidal suspensions and polymeric solutions will have different coating windows to Newtonian fluids. Figure 7 demonstrates how the I/C ratio and water/alcohol ratio influenced the coating windows of catalyst inks. As with MicrogravureTM, coating catalyst layers in a single pass creates drying issues resulting in surface cracks. The crack area increases with the coated layer thickness required to achieve Pt loading targets in heavy-duty PEMFC. Uniformity of the coated layer is easily affected by deviations in the coating parameters, thus high-precision equipment is crucial. Figure 8 compares the catalyst layers obtained through slot-die and spray-coating utilising two commercial catalysts, i.e. the Umicore Elyst Pt50 0550 (carbon black support) and the
Figure 5—Illustration of the slot-die coating process
Figure 6—Coating qualities obtained for slot-die coated catalysts layers (a) barring, (b) ribbing, (c) unstable meniscus and coating bead and (d) defectfree coating
Roll-to-roll coating methods for the manufacture of
Heraeus H2FC-50Pt-C700 (high surface area carbon support). For both catalysts, the spray coated counterpart resulted in uniform catalyst layers with no visible cracks, since spray coating is a multilayered process and drying occurs continuously throughout the coating process, thereby mitigating crack formation. The slot-die catalyst layers were deposited as a single layer and thus cracked during drying. The drying process of the wet coating significantly influences the morphology and microstructure of the catalyst layers. Solvent evaporation, sedimentation, agglomeration, assembly, packing, and internal particle migration due to concentration and temperature gradients, are the dominating factors that influence catalyst layer microstructure formation (Liu et al., 2023). Cracks have been reported to have a detrimental effect on PEMFC performance and have been linked to pin-hole formation in the membrane, local flooding, catalyst erosion, free radical generation, and increased catalyst layer resistance (Liu et al., 2024). Optimising the catalyst layer microstructure is crucial for PEMFC performance and is an important challenge that needs to be addressed in slot-die coating.
Influence of MicrogravureTM and slot-die coating on MEA performance
Beginning-of-life performances of MEAs fabricated using the MicrogravureTM, slot-die, and spray-coated GDEs were evaluated. Figure 9 (a) shows the polarisation curves of MEAs utilising the Umicore Elyst Pt50 0550 catalyst. The MEAs composed of the MicrogravureTM and spray-coated GDEs showed comparable performance with the MicrogravureTM having marginally better performance at 0.6V. However, at higher current densities, the spray-coated MEA showed improved performance, achieving higher peak power density. The improvement at higher current densities for the spray-coated MEA can be linked to the microstructure of the catalyst layer with fine cracks and craters in the MicrogravureTM catalyst layers, possibly leading to localised flooding and higher mass transport resistance. The performance metrics are summarised in Table 1. The slot-die coated MEA showed notably lower performance than the MicrogravureTM and spray-coated MEAs, which could be due to the highly cracked surface of the catalyst layers. Although the three MEAs had similar electrochemical surface areas (ECSA), the spray-coated MEA showed significantly better ORR activity indicating better Pt utilisation. Figure 9 (b) shows the polarisation curves of MEAs utilising the Heraeus H2FC-50Pt-C700 catalyst. Since the ink containing this catalyst did not coat with MicrogravureTM, no MEA result was available. MicrogravureTM coating of this catalyst will be explored in future work. The figure compares the performance of the slot-die MEA and the spray-coated MEA. As observed with the Umicore catalyst, the spray-coated MEA outperformed the slot-die MEA. In this
8—Micrographs of slot-die and spray-coated GDEs. Coatings with Umicore and Heraeus catalysts utilising carbon black and a high surface area carbon, respectively, are also compared
case, the spray-coated MEA exhibited a larger ECSA of 102.69 m2/ gPt compared to 71.39 m2/gPt, demonstrating that the uniform catalyst layer in spray-coating resulted in better Pt utilisation. Optimising drying conditions, especially for slot-die, is crucial for uniform crack-free catalyst layers and improved MEA performance. Although the spray-coated MEA with Umicore catalyst achieved the highest peak power density, the MicrogravureTM MEA with Umicore catalyst and the spray-coated MEA with Heraeus catalyst achieved comparable peak power densities at higher voltages, indicating less Pt degradation should occur during peak PEMFC operation.
Conclusions
MicrogravureTM and slot-die coating methods were evaluated for GDE fabrication, and preliminary MEA performance results indicate that MicrogravureTM has a strong potential for high-volume GDE production. Slot-die coated catalyst layers yielded significant cracking during drying, affecting the MEA performance. Although Microgravure catalyst layers exhibited cracking, it did so to a much lesser extent than slot-die coating. Addressing challenges related to catalyst layer drying is crucial to producing high-performance MEAs at high-volume scales. The use of catalyst inks with higher boiling point solvents and/or crack-inhibiting additives will be investigated in future work. Moreover, optimising other catalyst ink variables, such as the solvent composition, is crucial in MicrogravureTM and slot-die methods to ensure wettability and catalyst ink spreading on the GDL surface. Catalyst inks with high contact angles do not wet the GDL surface well and are prone to producing defective coatings. In slot-die, precise control of other coating variables such as web speed, ink dispensing rate, gap height, etc. is also crucial since deviations in any of these variables will lead to defective coatings. The MicrogravureTM method is generally simpler than slot-die, requiring control of fewer parameters. However, MicrogravureTM is restricted by a wet layer thickness maximum, which limits the scope of application of the method.
Acknowledgements
The authors are grateful to Mintek for the research facilities, equipment, and resources to conduct this work.
Figure 7—Coating windows for catalyst inks with different I/C and water/ alcohol ratios
Figure
Figure 9—Polarisation and power curves of MicrogravureTM, slot-die, and spray-coated MEA with (a) Umicore and (b) Heraeus catalysts. Cell temperature = 80°C, a/c: RH = 100/80, 2/2 bar backpressure, hydrogen/air flow rates = 0.43/2 nlpm. Cell fixture pressure at 4.8 bar
Table 1
Electrochemical properties of the MEAs evaluated MEA
SD – Heraeus
SC – Heraeus 945 563
References
Chang, Y.-R., Chang, H.-M., Lin, C.-F., Liu, T.-J, Wu, P.-Y., 2007. Three minimum wet thickness regions of slot die coatng. Journal of Colloid and Interface Science, vol. 308, pp. 222–230.
Creel, E.B., Tjiptowidjojo, K.A., Livingston, K., Randal Schunk, P., Bell, N., Serov, A., Wood, D. 2022. Slot-die-coating operability windows for polymer electrolyte membrane fuel cell cathode catalyst layers. Journal of Colloid and Interface Science, vol. 610, pp. 474–485.
Debe, M. 2012. Electrocatalyst approaches and challenges for automotive fuel cells. Nature, vol. 486, pp. 43–51.
Ding, X., Harris, T.A.L. 2017. Review on penetration and transport phenomena in porous media during slot die coating. Journal of Polymer Science, Part B: Polymer Physics, vol. 55, pp. 1669–1680.
Ding, X., Liu, J., Harris, T. 2016. A review of the operating limits in slot die coating processes. AIChE Journal, vol. 62, pp. 2508–2524.
Grand View Research. 2023. Proton Exchange Membrane Fuel Cell Market Size, Share & Trends Analysis Report By Type (High Temperature, Low Temperature), By Application (Stationary, Mobile), By Material, By Region, And Segment Forecasts, 2024 - 2030, s.l.: s.n.
Hasegawa, N., Kamiya, A., Matsunaga, T., Kitano, N., Harada, M. 2021. Analysis of crack formation during fuel cell catalyst ink drying process. Reduction of catalyst layer cracking by addition of high boiling point solvent. Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol. 628, p. 127153.
Hewson, R., Kapur, N., Gaskell, P. 2011. A two-scale model for discrete cell gravure roll coating. Chemical Engineering Science, vol. 66, pp. 3666–3674.
Higgens, B.G., Scriven, L.E. 1980. Capillary pressure and viscous pressure drop set bounds on coating bead operability. Chemical Engineering Science, vol. 35, no. 3, pp. 673–682.
Kapur, N., Hewson, R., Sleigh, P.A., Summers, J.L., Thompson, H.M., Abbott, S.J. 2011. A review of gravure coating systems. Convertech & e-Print, vol. 5, pp. 56–60
Koh, H.J., Kwon, I., Jung, H.W., Hyun, J.C. 2012. Operability window of slot coating using viscocapillary model for Carreau- type coating liquids. Korea-Australian Rheology Journal, vol. 24, no. 2, pp. 137–141.
Lee, S.H., Koh, H.J., Ryu, B.K., Kim, S.J., Jung, H.W., Hyun, J.C. 2011. Operability coating windows and frequency response in slot coating flows from a viscocapillary model. Chemical Engineering Science, vol. 66, no. 21, pp. 4953–4959.
Liu, P., Yang, D., Li, B., Kang, J., Zhang, C., Ming, P., Pan, X., Liu, H. 2024. Benzoic acid as additive: A route to inhibit the formation
of cracks in catalyst layer of proton exchange membrane fuel cells. Journal of Power Sources, vol. 591, p. 233817.
Liu, P., Yang, D., Li, B., Zhang, C., Ming, P. 2023. Recent progress of catalyst ink for roll-to-roll manufacturing paired with slot die coating for proton exchange membrane fuel cells. International Journal of Hydrogen Energy, vol. 48, pp. 19666–19685.
Mauger, S.A., Neyerlin, K.C., Yang-Neyerlin, A.C., More, K.L., Ulsch, M. 2018. Gravure Coating for Roll-to-Roll Manufacturing of Proton-Exchange-Membrane Fuel Cell Catalyst Layers. Journal of the Electrochemical Society, vol. 165, no. 11, pp. F1012–F1018.
Medina, S., Foster, J.G., Dzara, M.J., Wang, M., Ulsch, M., Mauger, S.A., Pylypenko, S., 2023. Multi-technique characterization of spray coated and roll-to-roll coated gas diffusion fuel cell electrodes. Journal of Power Sources, vol. 560, p. 232670.
Park, J., Kang, Z., Bender, G., Ulsch, M., Mauger, S.A. 2020. Roll- to-roll production of catalyst coated membranes for low- temperature electrolyzers. Journal of Power Sources, vol. 479, pp. 2–8.
Rashid, T., Liang, H.-L., Taimur, M., Chiodarelli, N., Khawaja, H., Edvardsen, K., de Volder, M. 2021. Roll to roll coating of carbon nanotube films for electro thermal heating. Cold Regions Science and Technology, vol. 182, p. 103210.
Sarka, F., Tobis, Z. 2022. Design Issues for Slot-Die Coating Heads - Case Study. IOP Conference Series: Materials Science and Engineering, vol. 1237, p. 012014.
Sharma, J., Lyu, X., Reshetenko, T., Polizos, G., Livingston, K., Li, J., Wood III, J.L., Seroc, A. 2022. Catalyst layer formulations for slot-die coating of PEM fuel cell electrodes. International Journal of Hydrogen Energy, vol. 47, pp. 35838–35859.
Stahler, M., Stahler, A., Scheppers, F., Carmo, M., Stolten, D. 2019. A completely slot die coated membrane electrode assembly. International Journal of Hydrogen Energy, vol. 44, pp. 7053–7058.
U.S Department of Energy, 2023. DOE Technical Targets for Polymer Electrolyte Membrane Fuel Cell Components. [Online] Available at: https://t.ly/Q0y0X [Accessed 16 09 2024].
Vak, D., Weerasinghe, H., Ramamurthy, J., Subbiah, J., Brown, M., Jones, D. 2016. Reverse gravure coating for roll-to-roll production of organic photovoltaics. Solar Energy Materials & Solar Cells, vol. 149, pp. 154–161.
Xie, M., Chu, T., Wang, T., Wam, K., Yang, D., Li, B. 2021. Preparartion, Performance and Challenges of Catalyst Layer for Proton Exchange Membrane Fuel Cell. Membranes, vol. 11, p. 879.
Yasui-Seiki Co. (USA). The Yasui Seiki "Micro Gravure" coating method u
NATIONAL & INTERNATIONAL ACTIVITIES
26-29 May 2025 — 9TH Sulphur and Sulphuric Acid Conference 2025
Protea Hotel Stellenbosch and Conference Centre, Stellenbosch
Contact: Gugu Charlie Tel: 011 538-0238
E-mail: gugu@saimm.co.za
Website: http://www.saimm.co.za
3-5 June 2025 — Introduction to Mining Hydrology and Mine Dewatering Design Workshop
Indaba Hotel, Spa and Conference Centre, Fourways
Contact: Gugu Charlie Tel: 011 538-0238
E-mail: gugu@saimm.co.za
Website: http://www.saimm.co.za
22-25 June 2025 — European Metallurgical Conference 2025 (EMC) Hamburg, Germany
Website: https://gdmb.de/home/
25-26 June 2025 — 4TH Digital Transformation in Mining Conference 2025
Putting digital technologies to work Glenburn Lodge and Spa, Muldersdrift
Contact: Gugu Charlie Tel: 011 538-0238
E-mail: gugu@saimm.co.za
Website: http://www.saimm.co.za
21-25 July 2025 — AfriRock 2025
Pioneering Progress: The Future of Rock Engineering Sun City, South Africa
Contact: Camielah Jardine
Tel: 011 538-0237
E-mail: camielah@saimm.co.za
Website: http://www.saimm.co.za
18-19 August 2025 — 13TH International Heavy Minerals Conference 2025
Sun City Resort, Rustenburg, South Africa
Contact: Gugu Charlie Tel: 011 538-0238
E-mail: gugu@saimm.co.za
Website: http://www.saimm.co.za
1-3 October 2025 — International Mineral Asset Valuation Conference 2025
Navigating Mineral Asset Valuations in an Uncertain Future
The Maslow Hotel, Sandton
Contact: Gugu Charlie Tel: 011 538-0238
E-mail: gugu@saimm.co.za
Website: http://www.saimm.co.za
13-16 October 2025 — Geometallurgy Conference 2025
Future-Ready Geometallurgy: Trusted Data, Advanced Tools, Smarter Decisions
Glenburn Lodge and Spa, Muldersdrift
Contact: Gugu Charlie Tel: 011 538-0238
E-mail: gugu@saimm.co.za
Website: http://www.saimm.co.za
21 October 2025 — 21ST Annual Student Colloquium 2025
Innovation as a Catalyst for Process Development and Job Creation
Mintek, Randburg, Johannesburg
Contact: Gugu Charlie Tel: 011 538-0238
E-mail: gugu@saimm.co.za
Website: http://www.saimm.co.za
27-29 October 2025 — 9TH International PGM Conference 2025
PGM - Enabling a cleaner world
Sun City, Rustenburg, South Africa
Contact: Gugu Charlie Tel: 011 538-0238
E-mail: gugu@saimm.co.za
Website: http://www.saimm.co.za
30-31 October 2025 — 8TH Young Professionals Conference 2025
Mintek, Randburg, Johannesburg
Contact: Gugu Charlie Tel: 011 538-0238
E-mail: gugu@saimm.co.za
Website: http://www.saimm.co.za
16-22 November 2025 — The 12TH International Copper Conference (Copper 2025)
Phoenix, Arizona, USA
Website: https://www.extractionmeeting.org/ Extraction2025/Extraction2025/Copper2025/default.aspx
19-21 November 2025 — MineSafe Hybrid Conference 2025
Safe Mines, Healthy Lives, and Sustainable Futures
Emperors Palace Convention Centre
Contact: Camielah Jardine
Tel: 011 538-0237
E-mail: camielah@saimm.co.za
Website: http://www.saimm.co.za
Company affiliates
The following organizations have been admitted to the Institute as Company Affiliates
acQuire Technology Solutions
AECI Mining Chemicals, a division of AECI Mining Ltd
African Pegmatite
Allied Furnace Consultants
AMIRA International Africa (Pty) Ltd
Anglogold Ashanti Ltd
Anton Paar Southern Africa
Arcus Gibb (Pty) Ltd
Becker Mining (Pty) Ltd
Bluhm Burton Engineering Pty Ltd
Buraaq mining Services (Pty) Ltd
Caledonia Mining South Africa
CARBOCRAFT
Castle Lead Works
CIGroup ACE Pty Ltd
DDP Specialty Products South Africa (Pty) Ltd
De-Tect Unit Inspection (Pty) Ltd
Digby Wells and Associates
EHL Consulting Engineers (Pty) Ltd
Elbroc Mining Products (Pty) Ltd
EPIROC South Africa (Pty) Ltd
Ex Mente Technologies (Pty) Ltd
Exxaro Resources Limited
FLSmidth Minerals (Pty) Ltd
G H H Mining Machines (Pty) Ltd
Geobrugg Southern Africa (Pty) Ltd
Glencore
Gravitas Minerals (Pty) Ltd
Hatch (Pty) Ltd
Herrenknecht AG
Impala Platinum Holdings Limited
IMS Engineering (Pty) Ltd
Ingwenya Mineral Processing
Ivanhoe Mines SA
Malvern Panalytical (Pty) Ltd
Maptek (Pty) Ltd
Mech-Industries (Pty) Ltd
Micromine Africa (Pty) Ltd
Minearc South Africa (Pty) Ltd
Minerals Council of South Africa
MineRP Holding (Pty) Ltd
Mining Projection Concepts (Pty) Ltd
Mintek
MLB Investments CC
Modular Mining Systems Africa (Pty) Ltd
Murray & Roberts Cementation (Pty) Ltd
Paterson & Cooke Consulting Engineers (Pty) Ltd
Pump and Abrasion Technologies (PTY) Ltd
Redpath Mining (South Africa) (Pty) Ltd
Rosond (Pty) Ltd
Roytec Global (Pty) Ltd
Rustenburg Platinum Mines Limited - Union
Salene Mining (Pty) Ltd
Schauenburg (Pty) Ltd
Sebotka (Pty) Ltd
SENET (Pty) Ltd
Sibanye Gold Limited
Sound Mining Solution (Pty) Ltd
SRK Consulting SA (Pty) Ltd
StageFright Edutainment
Tomra (Pty) Ltd
Trans-Caledon Tunnel Authority
Ukwazi Mining Solutions (Pty) Ltd
VBKOM Consulting Engineers
Weir Minerals Africa
ZUTARI (Pty) Ltd
Experts in Slurry Pipelines, Mine Backfilling, Mine Waste Systems, and Mineral Processing
With over 30 years of experience and 9 offices worldwide, our global presence allows us to offer local knowledge to our clients, paired with international expertise.
Slurry Pipelines
As a global leader in slurry pipeline systems our engineers have experience in all aspects of slurry pipeline system design and operation, ranging from laboratory test work, hydraulic analysis, pipeline routing, and pipeline construction, to the detailed design and engineering of pump, valve, and choke stations.
Mine Backfill
Since 1991, our engineers have solved some of the most complex backfill challenges while testing and designing hydraulic, paste and cemented aggregate fill projects all over the world. We pioneered the development of full flow technology to keep backfill reticulation systems fully pressurised to manage energy and reduce pipeline wear.
Our global backfill group, made up of engineers, operators, technologists, and specialists, allows us to coordinate expertise and resources from our offices to deliver backfill projects through detail design, construction, commissioning and ongoing operational support. Our expertise includes the detailed understanding of the backfill requirements to support the mining method, cement technology to optimise the binder selection, process development and specialist engineering services.
Mine Waste
Paterson & Cooke’s holistic solutions balance many competing requirements such as geotechnical stability, geochemical stability, water conservation, social license to operate, people at risk, emissions, risk to operations, and cost.
Our process starts with understanding corporate and mine site requirements in conjunction with the client and the geotechnical engineer.
Through this collaborative process, we develop a technology roadmap that addresses each client’s unique challenges. The solutions that are developed are site specific and focus on each of the processing, transportation, and deposition challenges.
Mineral Processing
With declining feed grades, increasingly challenging and problematic ores being processed, and a drive for more sustainable mining practices, the modern process plant needs to adapt. Often this means adopting newer and novel technology or modifying and/or optimising the process flowsheet. From early-stage conceptual studies through to construction, commissioning and training, we provide solutions every step of the way.
Laboratory Services
We have comprehensive laboratories in the countries in which we operate that provide a wide range of metallurgical, process and bulk materials handling test work to support our engineering designs, as well as providing flow sheet development and pilot plant campaigns. Our testing capabilities include processing run of mine ore to generate concentrate and flotation tailings as well as extensive crushing, milling, flotation, dewatering, geochemical and mineral analysis.
Geometallurgy Conference 2025
Future-Ready Geometallurgy: Trusted Data, Advanced Tools, Smarter Decisions
13 October 2025 Technical Workshops
14-15 October 2025 Conference
16 October 2025 Technical Visits
Venue: Glenburn Lodge and Spa, Muldersdrift
(1-hour drive from O.R. Tambo International Airport, Johannesburg, South Africa)
ECSA Validated CPD Activity, Credits = 0.1 points per hour attended.
BACKGROUND
Ore heterogeneity and metallurgical complexity continue to rise, while sustainability occupies an ever-prominent role. Geometallurgy offers a refined, multidisciplinary pathway to improve the value of ore deposits.
Future-Ready Geometallurgy examines strategies to generate trustworthy data through the steady application of machine learning, novel sensors, digital twins, and Industry 4.0 technologies in mining. Reliance on dependable data and innovative tools empowers the industry to make more timely and discerning decisions for mine planning and metallurgical plant optimisation.
Following the success of past SAIMM Geometallurgy conferences, the third conference in this series provides a platform to explore the latest progress in geometallurgy and to celebrate the success of its integration into the mining value chain.
CONFERENCE THEMES
• Sampling and sensor-based core logging
• Practical tools and methodologies for ore characterisation and testing
• Standardised methods for reporting to generate trusted data
• Geostatistics, data integration, smart workflows, and modelling
• Incorporation of predictive mining, processing, and environmental models into resource modelling
• Project and operational geometallurgy case studies
KEY DATES
• 17 June 2025 - Submission of short abstracts
• 29 July 2025 - Submission of extended abstracts for peer review
• 13 October 2025 – Technical workshops
• 14-15 October 2025 - Conference
• 16 October 2025 - Technical visits
FOR FURTHER INFORMATION CONTACT:
Gugu Charlie, Conferences and Events Co-ordinator E-mail: gugu@saimm.co.za
Keynote Speakers
Prof. Julian M. Ortiz Ph.D., P.Eng.
Mark Cutifani / Anglo American Chair in Mining Innovation, Camborne School of Mines | University of Exeter
Wendy Ware
Geomet Advisor & Business Development Manager, Datarock, Australia
Call for Abstracts
Prospective authors are invited to submit short abstracts of not more than 500 words to: Conferences and Events Co-ordinator, Gugu Charlie, e-mail: gugu@saimm.co.za
Authors of accepted abstracts will be required to submit extended abstracts for peer review and publication in the conference proceedings. Following the conference, full papers can be submitted to the journal for peer review and publication in an SAIMM Journal Special Issue.